Yudha P. Budiman†
*a,
Muhamad R. S. Sidik†
a,
Muhamad Diki Permana
bc,
Kansy Haikala,
Iis I. Widiyowatia,
Yessi Permana
d,
Ubed S. F. Arrozi
e,
Wirawan Ciptonugroho
f,
Tri Mayanti
a,
Allyn P. Sulaeman
a,
Juliandri
a and
Witri Wahyu Lestari
f
aDepartment of Chemistry, Faculty of Mathematics and Natural Sciences, Universitas Padjadjaran, 45363 Sumedang, Indonesia. E-mail: y.p.budiman@unpad.ac.id
bSpecial Educational Program for Green Energy Conversion Science and Technology, Integrated Graduate School of Medicine, Engineering, and Agricultural Sciences, University of Yamanashi, Kofu 400-8511, Japan
cCenter for Crystal Science and Technology, University of Yamanashi, Kofu 400-8511, Japan
dInorganic and Physical Chemistry Research Division, Faculty of Mathematics and Natural Sciences, Institut Teknologi Bandung, Bandung 40132, Indonesia
eDepartment of Chemistry, Faculty of Mathematics and Natural Sciences, State University of Malang, 65145, Malang, Indonesia
fDepartment of Chemistry, Faculty of Mathematics and Natural Sciences, Universitas Sebelas Maret, Surakarta, 57126, Indonesia
First published on 21st August 2025
This study investigates the catalytic potential of copper-based metal–organic frameworks (Cu-MOFs), specifically HKUST-1 and CuBDC, for the protodeboronation of aryl boronic acids. Protodeboronation, was explored under various bases, atmospheres, and substrates. Optimal conditions using K2CO3 as the base and an oxygen atmosphere yielded up to 98% product with HKUST-1. While CuBDC also exhibited catalytic activity, its yields were slightly lower under identical conditions. Substrate size and substituent effects played a crucial role, with bulkier substrates favoring higher yields. Recyclability tests confirmed that both Cu-MOFs retained catalytic activity over three cycles, despite some structural changes. These findings demonstrate Cu-MOFs as promising heterogeneous catalysts for controlled protodeboronation.
![]() | ||
Scheme 1 (a) Boronic acids utilized in coupling reactions and (b) protodeboronation of boronic acids. |
Despite their utility, these reactions often encounter a competing pathway known as protodeboronation, where the carbon–boron bond is replaced with a carbon–hydrogen bond (Scheme 1b).18 This side reaction is typically undesirable, as it reduces the yield of the desired coupling product, thereby impacting the overall efficiency of the reaction.19
The first detailed study of this process dates back to 1930, when Ainley and Challenger observed that phenylboronic acid in water at 140–150 °C formed benzene after 40 h in the presence of stoichiometric amounts of salts (ZnCl2, CrBr3, CuSO4).20 Its synthetic potential was later explored by Brown and Zweifel in 1961.21 Since then, protodeboronation has evolved from being viewed as a side reaction to becoming an intentional step in certain synthetic pathways. For instance, Lai et al. specifically studied a palladium-catalyzed protodeboronation of arylboronic acids under basic conditions and a nitrogen atmosphere.22 In 2011, Elford et al. utilized the protodeboronation of boronic acid esters as a pivotal step in the synthesis of the natural product (+)-erogorgiaene.23 In the same year, Veguillas et al. demonstrated the synthesis of quinonyl boronic acid derivatives, where the boronic acid group played a critical role in initiating Friedel–Crafts alkylation via protodeboronation.24 Further developments were reported in 2013 by Lee et al., who synthesized o- and m-phenols using phenylboronic acid as a precursor, with the boronic acid group serving as a temporary blocking or directing group before being removed through protodeboronation.25 In 2014, Lozada et al. studied a variety of electron-deficient and heteroarylboronic acids subjected to protodeboronation under basic conditions.26 More recently, Budiman et al. reviewed the influence of o-fluoro substituents on the reactivity of arylboronic acids in basic aqueous conditions, demonstrating how functional groups can enhance the likelihood of protodeboronation.27 These milestones underscore the growing importance of understanding and controlling protodeboronation, not only to minimize its occurrence as an undesirable side reaction but also to harness its potential as a deliberate tool in organic synthesis.
To the best of our knowledge, the use of metal catalysts for protodeboronation has been extensively explored with metal salt systems such as copper,28 silver,29 iridium,30 and cobalt,31 among others (Scheme 2). In contrast, metal–organic frameworks (MOFs) present an emerging class of recyclable heterogeneous catalysts with significant potential, yet their application in protodeboronation reactions remains unexplored. The combination of hybrid inorganic–organic building blocks forming coordinated porous frameworks and the intrinsic rigidity of these frameworks imparts a wide range of functionalities with desirable properties.32 These features include active metal centers, adjustable pore sizes with high porosity, large surface areas, building block versatility, and flexible topological designs.33 MOFs have attracted considerable attention in organic synthesis due to their role as versatile heterogeneous catalysts with capabilities in catalyzing reactions such as Friedel–Crafts alkylation,34,35 Diels–Alder,36 cyclization,37 Michael addition,38 Claisen–Schmidt condensation,39 and CO2 cycloaddition.40 The adaptable structures of MOFs, both pre- and post-synthesis, enable precise control over their secondary building units (SBUs), pore functionalization, and the abundance of active sites resulting from higher metal content.41 These features have the potential to tailor the type of MOF catalyst needed for certain reactions. It is also important to underline that due to the inherent complexity of MOF structures, fully understanding the mechanisms of MOF catalytic systems is a challenging task, as both the metal nodes and organic linkers, as well as pore interactions, could play a role in providing catalytic sites.42
Building on our previous work optimizing HKUST-1 as a catalyst for the homocoupling of arylboronic acids,43 we now report the protodeboronation of arylboronic acids catalyzed by copper-based MOFs. Specifically, we employed HKUST-1 and CuBDC as catalysts. Both MOFs feature copper as the central metal site; however, they differ in their organic linkers. HKUST-1 utilizes benzenetricarboxylic acid (H3BTC),44 forming a three-dimensional framework, while CuBDC incorporates benzenedicarboxylic acid (H3BDC),45 resulting in a two-dimensional structure. This study explores the catalytic potential of these frameworks in protodeboronation reactions of arylboronic acids.
For reactions under controlled atmospheres, N2 or O2 was introduced by inflating a balloon attached to the syringe with the respective gas, while reactions in open air were performed by removing the septum from the Schlenk tube. The reaction was carried out at 70 °C for 1.5 h.
The reaction products were extracted using brine water (saturated NaCl solution, 2 mL) and ethyl acetate (2 mL), and washed with ethyl acetate up to three times. The ethyl acetate layer was accumulated, dried over anhydrous Na2SO4, filtered, and purified via column chromatography using n-hexane as the eluent to isolate the protodeboronation product. The extracted ethyl acetate layer containing the protodeboronation products were directly transferred to a vial after adding an equimolar amount of mesitylene (0.4 mmol) before being analyzed by GC-MS (Agilent 7890A) for its GC yield. Mesitylene was used as an internal standard to compare the GC response of the obtained products.
For catalyst recycling procedure, after extraction, the aqueous layer was centrifuged up to three times. The water was then decanted to separate it from the precipitated catalyst, which was recovered for reuse in another run.
Entry | Catalyst | Base | Base equivalent | Yield (%) |
---|---|---|---|---|
a Reaction conditions: 4-tert-butylphenylboronic acid (0.4 mmol), Cu-MOFs (10 mol%), EtOH/H2O (1 mL/1 mL), O2, 70 °C, 1.5 h. GC yields were reported with mesitylene as an internal standard. | ||||
1 | — | — | — | — |
2 | — | K2CO3 | 1.0 | 31 |
3 | HKUST-1 | — | 12 | |
4 | KHCO3 | 1.0 | 29 | |
5 | CH3COOK | 1.0 | 36 | |
6 | Na2CO3 | 1.0 | 54 | |
7 | NEt3 | 1.0 | 78 | |
8 | CsF | 1.0 | 65 | |
9 | K2CO3 | 1.0 | 98 | |
10 | K2CO3 | 0.5 | 69 | |
11 | K2CO3 | 0.2 | 46 | |
12 | CuBDC | — | — | 19 |
13 | KHCO3 | 1.0 | 38 | |
14 | CH3COOK | 1.0 | 53 | |
15 | Na2CO3 | 1.0 | 47 | |
16 | NEt3 | 1.0 | 83 | |
17 | CsF | 1.0 | 51 | |
18 | K2CO3 | 1.0 | 95 | |
19 | K2CO3 | 0.5 | 52 | |
20 | K2CO3 | 0.2 | 47 |
To further investigate the impact of the different Cu-MOF catalysts used in the model, a series of bases were tested using both Cu-MOF catalysts. The choice of base had significant impact on the outcome, as the HKUST-1-catalyzed reactions with bases such as KHCO3, CH3COOK, Na2CO3, NEt3, and K2CO3 produced low to excellent yields (Table 1, entries 4–9). Similarly, the CuBDC-catalyzed reactions yielded comparable results to HKUST-1, with slight variations in yield on the same base systems (Table 1, entries 13–18). A slight modification was explored by reducing the equivalent amount of one of the base variants to 0.5 and 0.2 equivalents. The results were consistent for both Cu-MOF catalysts, showing that a decrease in the amount of base led to a corresponding reduction in the reaction yield (Table 1, entries 10, 11 and 19, 20). Thus, we selected K2CO3 as the base for the next section of this research.
Entry | Catalyst | Atmosphere | Yieldb (%) |
---|---|---|---|
a Reaction conditions: 4-tert-butylphenylboronic acid (0.4 mmol), Cu-MOFs (10 mol%), K2CO3 (1 eq.) EtOH/H2O (1 mL/1 mL), 70 °C, 1.5 h.b GC yields were reported with mesitylene as an internal standard. | |||
1 | HKUST-1 | Air | 65 |
2 | O2 | 98 | |
3 | N2 | 54 | |
4 | CuBDC | Air | 76 |
5 | O2 | 95 | |
6 | N2 | 43 |
Entry | Boronic reagents | Products | Yieldb (%) | Conversionc (%) |
---|---|---|---|---|
a Reaction conditions: arylboronic acid (0.4 mmol), HKUST-1 (10 mol%), K2CO3 (1 eq.) EtOH/H2O (1 mL/1 mL), O2, 70 °C, 1.5 h.b GC yields were reported with mesitylene as an internal standard.c Conversion based on remaining boronic reagents detected in GC.d Using CuBDC (10 mol%).e Due to high volatility of fluorobenzene (2f), the reaction proceeds at 40 °C.f Isolated yield after column chromatography. | ||||
1 | ![]() |
![]() |
47(20)d | 100 |
2 | ![]() |
![]() |
34(15)d | 100 |
3 | ![]() |
![]() |
29 | 100 |
4 | ![]() |
![]() |
60 | 100 |
5 | ![]() |
![]() |
57 | 100 |
6e | ![]() |
![]() |
27(24)d | 100 |
7 | ![]() |
![]() |
68 | 100 |
8 | ![]() |
![]() |
97(95)f | 100 |
9 | ![]() |
![]() |
98(95)d | 100 |
10 | ![]() |
![]() |
60(54)f | 100 |
11 | ![]() |
![]() |
41 | 100 |
12 | ![]() |
![]() |
trace | 100 |
13 | ![]() |
![]() |
64(62)f | 100 |
14 | ![]() |
![]() |
Trace | 100 |
A similar effect was observed for bulkier substrates, including biphenyl-, 4-tert-butylphenyl-, and benzo[b]thien-2-ylboronic acids (1h, 1i, 1j), which yielded significant amounts of the corresponding protodeboronated products (Table 3, entries 8–10). Among electron-withdrawing substrates, p-fluorophenylboronic acid (1f) and p-bromophenylboronic acid (1g) also followed this trend, with the former yielding noticeably lower amounts, likely due to its smaller molecular size and lower reaction temperature (Table 3, entries 6 and 7).
To further investigate this behavior, we compared HKUST-1 with CuBDC as a catalyst for selected substrates (Table 3, entries 1, 2, 6, and 9). Interestingly, p-tolyl, m-tolyl, and p-fluorophenylboronic acids (1a, 1b, 1c) gave lower yields with CuBDC, while 4-tert-butylphenylboronic acid (1i) showed no significant difference between the two catalysts. An additional investigation was conducted to verify whether the small molecule was adsorbed into HKUST-1. To this end, 0.4 mmol of toluene was added as a substrate to 1 mL of ethanol containing 10 mol% of HKUST-1, and the mixture was stirred overnight. Gas chromatography (GC) analysis showed no detectable peak corresponding to toluene, confirming its adsorption into the MOF (Fig. S2.19). This supports the hypothesis that small molecules are more likely to be adsorbed within the MOF framework. Furthermore, steric effects play a role, as o-substituted substrates generally yielded lower amounts of product than their p- and m-substituted counterparts.
In addition to arylboronic acids, we examined heteroarylboronic acids and alkenyl- and alkyl-based boronic acids to expand the substrate scope. Benzo[b]thien-2-ylboronic acid (1j) underwent successful protodeboronation with HKUST-1 (Table 3, entry 10). Meanwhile, trans-2-phenylvinylboronic acid (1k) gave a moderate yield of the protodeboronated product (2k), whereas the protodeboronation of phenethylboronic acid (1l) was not detected, and interestingly it shows a minor hydroxylation product was observed. We have previously reported that HKUST-1 can promote the hydroxylation of electron-rich aryl–B(OH)2 under moist conditions in DMF, in the presence of O2 and notably in the absence of base.43 We believe that, in addition to minor formation of the hydroxylated byproduct of 1l, the majority of the resulting protodeboronation products may have remained adsorbed within the porous HKUST-1 framework, making them undetectable by standard GC-MS analysis.52–54 To verify this, we performed the reaction under standard conditions using ethylbenzene. After 1.5 hours, and with mesitylene as an internal standard, no ethylbenzene was detected by GC-MS, indicating its complete adsorption by HKUST-1.
Finally, we evaluated boronic acid pinacol esters (–Bpin) as substrates (Table 3, entries 13 and 14). A bulkier substrate, 2-naphthyl-Bpin (1m), underwent protodeboronation with moderate efficiency (2m). In contrast, a smaller substrate, m-tolyl-Bpin (1n), yielded only trace amounts of toluene (2a). Two key factors contribute to the absence of detectable protodeboronation products of m-tolyl-Bpin (1n) in the GC-MS analysis: (1) adsorption of toluene by HKUST-1, which reduces its volatility and recovery during analysis;52–54 (2) in situ formation of a boron–alkoxide adduct, such as [tolylBpin(OEt)]−, which is not amenable to GC-MS detection due to its ionic and non-volatile nature.
The latter has been demonstrated by Fernández and co-workers,57 who showed that –Bpin species can readily form Lewis acid–base adducts with alkoxide anions generated in situ from the combination of carbonate bases and alcohols. These adducts are typically non-volatile and thus remain undetected under standard GC-MS conditions.
Lloyd-Jones and co-workers have previously reported that the protodeboronation of electron-rich aryl boronic acids proceeds via a tetra-coordinate organoboronate intermediate formed through coordination with hydroxide.58 This is followed by a rate-limiting proton transfer from a water molecule, resulting in C–B bond cleavage. Therefore, in our case (Table 3, entry 14), the sluggish reactivity of m-tolyl-Bpin is likely halted at the stage of a stable boron–alkoxide adduct, which is undetectable by GC-MS.
To verify this, we conducted a control experiment in which m-tolyl-Bpin was treated with K2CO3 in an ethanol/water (1:
1) mixture and stirred at 75 °C for 1.5 hours in the absence of the HKUST-1 catalyst. Under these conditions, and using a stoichiometric amount of mesitylene as an internal standard, neither the starting material nor the protodeboronation product was detected by GC-MS. It is also important to note that electron-rich aryl boronates generally exhibit greater resistance to protodeboronation than their electron-deficient counterparts. In a subsequent study,59 the same group investigated electron-deficient aryl boronates, such as 2,6-difluorophenyl derivatives, and found that Bpin esters undergo the slowest protodeboronation among various boronate esters, including those derived from catechol, ethylene glycol, 1,3-propanediol, 2-hydroxymethyl-2-methyl-1,3-propanediol, neopentylglycol, and 2,4-dimethylpentan-2,4-diol. Thus, the protodeboronation of aryl-Bpin derivatives is generally more difficult than that of their corresponding boronic acid (–B(OH)2) analogues.
We propose that this protodeboronation reaction proceeds via a bimetallic pathway (Scheme 3).43,53,60 Initially, oxidation of complex A with oxygen generates Cu(III) species, which then react with Ar–B(OH)2, H2O, and K2CO3 to form intermediate B. This is followed by reductive elimination of the binuclear Cu(III) complex, producing the protodeboronated product and regenerating the Cu(II) catalyst in the presence of H2O, thereby sustaining the catalytic cycle.
To further investigate this, we conducted atomic absorption spectroscopy (AAS) analysis on the reaction mixture to assess potential copper leaching (see S4 SI). The analysis detected 0.081 ppm of copper, corresponding to only 0.0024% of the initial copper content, confirming that the MOF structure underwent structural changes, releasing a minimal amount of copper into the reaction medium. Nevertheless, the catalyst remained catalytically active for protodeboronation.
Structural changes of MOFs under basic conditions have been previously reported by Yuan et al. who observed that MOFs composed of high-valency metal ions and carboxylate-based ligands are particularly susceptible to degradation in alkaline environments.61 This instability arises from the low pKa of carboxylic acids, which, despite forming strong coordination bonds with high-valency metal ions, render these MOFs highly stable in acidic conditions but less so in basic media. In contrast, MOFs with high pKa ligands, such as azolate-based ligands paired with low-valency metal ions, are better suited for alkaline environments.
Footnote |
† These authors contributed equally. |
This journal is © The Royal Society of Chemistry 2025 |