The development of efficient, cost-effective and recyclable catalysts is a constant challenge for industries working with chemical transformations ^{[1, 2]}, and transition metal complexes with atomically-dispersed active centers have become one of the most intensively studied systems for this purpose. These homogeneous catalysts have well-defined structures, allowing for the fine-tuning of their electronic and steric properties [2−4]. However, a tedious ligand screening process is typically necessary to optimize the stability, activity and selectivity of a homogenous catalyst. Moreover, this class of catalysts has the disadvantage of not being easily separated from the reaction mixture, presenting more than a few engineering issues ^{[5, 6]}. Heterogeneous catalysts are a promising means of addressing the separation issue ^[7]. Unfortunately, intrinsically heterogeneous catalysts, such as supported metal nanoparticles, are usually less active than their homogeneous analogues. Heterogenization of homogeneous catalysts through chemical bonding on various supports, such as polymers or oxides, is an alternative strategy to mitigate the separation problem while maintaining high catalytic activity. In such cases, however, multistep synthetic procedures are normally required, limiting their large-scale applications [8−12]. Thus, it would be highly desirable to develop a simple and efficient synthetic route for the preparation of materials combining the advantages of homogeneous and heterogeneous catalysts.

We report herein the effective one-step synthesis of a novel nanoparticulate polyacetylene-supported Pd(II) catalyst, denoted as NP-Pd(II), that functions as a highly efficient and recyclable catalyst for the Suzuki-Miyaura cross coupling reaction. This material was obtained by simply treating an aqueous solution of PdCl₄²⁻ with acetylene under ambient conditions. Systematic characterization using electron microscopy, Fourier transform infrared (FT-IR) spectroscopy, Raman spectroscopy, X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS) and extended X-ray absorption fine structure (EXAFS) analysis demonstrated that the resulting NP-Pd(II) contained micron-sized aggregates of fine nanoparticles composed of polyacetylene-supported Pd(II) complexes. The homogeneous distribution of Pd(II) along the polyacetylene backbone, combined with the tendency of the material to aggregate, makes NP-Pd(II) an ideal catalyst, with the advantages of both homogeneous and heterogeneous catalysts ^{[13, 14]}. The prepared NP-Pd(II) was both air- and moisture-stable and exhibited high catalytic activity comparable to that of homogenous catalysts and better recyclability than carbon-supported Pd catalysts, especially in aqueous systems.

Because the Suzuki reaction is one of the most efficient methods of constructing biaryl and substituted aromatic molecules, its impact on both academic and industrial research has been immense [8,15−21]. To date, many Pd complexes have been applied as homogenous catalysts in the Suzuki reaction and the continual search for effective ligands has been one of the most fascinating aspects of this field. As an example, significant success has been achieved in using phosphine ligands, especially in conjunction with low-reactivity substrates [22−24]. Investigations of other ligands, such as N-heterocyclic carbenes [25−27], amines ^{[28, 29]} and cyclic compounds ^[30], have also been widely reported in the literature. The demonstration of polyacetylene as an effective ligand for Pd(II) in this work further expands the range of ligands applicable to efficient Suzuki coupling catalysts. More importantly, the facile, one-step, readily scaled-up synthesis of NP-Pd(II) makes this heterogeneous Pd catalyst highly desirable for practical applications.

A H₂PdCl₄ aqueous solution(20 mmol/L) was prepared by reacting PdCl₂ with concentrated aqueous HCl in stoichiometric amounts at 70 °C for 1 h. A suitable quantity of this aqueous solution was subsequently transferred to a glass pressure vessel and acetylene was introduced with stirring at ambient temperature (25 °C) while the relative pressure was maintained at 0.3 atm for 15 min. On contact with the acetylene gas, the H₂PdCl₄ solution immediately became turbid. The resulting NP-Pd(II) was readily separated by centrifuging this colloidal suspension and the powder thus obtained was washed twice with H₂O and ethanol, followed by further centrifugation. Finally, the NP-Pd(II) was dried under vacuum at 40 °C overnight. Inductively coupled plasma (ICP) analysis of the resulting dry NP-Pd(II) confirmed almost complete conversion of the original PdCl₄²⁻ to NP-Pd(II). This synthetic procedure was readily scaled up. As shown in Fig. S1, 1 g of NP-Pd(II) was successfully obtained from a larger scale reaction performed in a 250-mL flask. Synthesis using an acetylene pressure of 1 atm rather than 0.3 atm gave a material termed NP-Pd(II)-1atm that was found to have a similar composition and morphology (Fig. S2) to the NP-Pd(II), and to generate similar XPS (Fig. S3) and FT-IR (Fig. S4(a)) spectra as well as exhibiting the same catalytic performance (Fig. S4(b)).

In a glass vessel, 15 mL H₂O, 212 mg Na₂CO₃, 183 mg phenylboronic acid and 112 μL iodobenzene were added. The vessel was then placed into a pre-heated water bath and the reaction was initiated by adding 0.24 mg NP-Pd(II). After the reaction was complete, diethyl ether (3×3 mL) was added to extract the products for gas chromatographic (GC) analysis. The products, still in diethyl ether, were analyzed by GC-mass spectrometry (MS) by comparison with standard samples of the expected products, and the extent of conversion was determined using n-heptane as an internal standard. With the exception of p-iodoanisole, which gave a small amount of biphenyl as by-product, most substrates exhibited complete selectivity for the desired product. The turnover frequency (TOF) of each reaction was typically calculated as the substrate turnover per mole of Pd per hour at a given temperature. To compare the results of the present study with results previously reported in the literature, the trials used to determine TOF values were conducted under similar conditions in terms of solvents and temperature. In recycling tests, the diethyl ether layer was carefully removed and fresh substrates were then added for further reaction under the same conditions.

The EXAFS studies of NP-Pd(II) were conducted at the 01C1 beamline of the National Synchrotron Radiation Research Center (NSRRC) at Hsinchu, Taiwan. A standard fluorescence configuration was employed, using an in-line metal reference foil for energy calibration. The NP-Pd(II) was dissolved in approximately 2 mL ethanol prior to analysis. Data processing and fitting were performed using the commercially available WinXAS and FEFF8 software packages. The resulting XAFS spectrum in k-space is shown in Fig. S5, over the range of 3 to 12 Å. EXAFS fitting was performed only over the region of interest (1.28-2.40 Å). More details of the fitting method applied to the Fourier-transformed EXAFS spectrum are provided in Supporting Information.

As noted, a typical synthesis of NP-Pd(II) was carried out by treating an aqueous solution of PdCl₄²⁻ with acetylene at room temperature. After collecting the product by centrifugation and washing with H₂O and ethanol, the NP-Pd(II) was obtained in the form of a brown-red powder (Supporting Information). The morphology of the NP-Pd(II) was characterized using transmission electron microscopy (TEM), and the TEM image in Fig. 1(a) shows that the obtained NP-Pd(II) was primarily in the form of spherical particles with an average diameter of approximately 100 nm. A high-magnification TEM image (Fig. 1(b)) further reveals that the as-prepared NP-Pd(II) particles were actually aggregates of very fine nanoparticles (FNPs) with diameters of 2-3 nm. Interestingly, when polyvinylpyrrolidone (PVP, 30K) was introduced to an aqueous dispersion of NP-Pd(II) with vigorous stirring, the aggregated particles were disassembled into isolated FNPs with a mean particle size of 2.8 ± 0.4 nm (Fig. 1(c) and (d)). This disassembly process suggests that the interaction between the FNP subcomponents was weak. The aggregation of FNPs into much larger NP-Pd(II) particles in water could be driven by their hydrophobic nature and the associated minimization of the total surface energy. Together with its aggregating nature, the hydrophobic nature of NP-Pd(II) allows the catalyst to be readily separated from water via either centrifugation or static settlement. Moreover, the synthesis of NP-Pd(II) may be readily scaled up to allow for mass production as a result of the one-step process and mild reaction conditions. As noted above, in our laboratory, NP-Pd(II) was successfully prepared on the gram scale using the newly developed synthetic method (Fig. S1).

Fig. 1. TEM images of NP-Pd (II) at low (a) and high (b) magnifications and after treatment with PVP (c).(d) is the size distribution of the fine nanoparticles of NP-Pd (II).

Elemental analyses using ICP-MS and a CHN analyzer determined that the Pd, C, H and Cl proportions in the NP-Pd(II) were 44.9%, 22.0%, 2.4% and 30.7%, respectively, thus the Pd:C:Cl ratio was approximately 1.0:4.3:2.1. Despite the high percentage of Pd, the XRD pattern of NP-Pd(II) displays no obvious diffraction peaks, indicating the absence of metallic Pd nanoparticles and the amorphous nature of the NP-Pd(II) (Fig. 2(a)). However, the electron beam-induced formation of metallic Pd nanoparticles was observed during TEM measurements, and so XPS was used to determine the oxidation state of Pd in the NP-Pd(II). Fig. 2(b) shows a Pd 3d doublet at binding energies of 341.7 and 336.6 eV, corresponding to 3d_3/2 and 3d_5/2, respectively. The binding energy of the Pd 3d_5/2 peak with good symmetry at 336.6 eV was significantly higher than that of Pd foil (ca. 335 eV), but close to that of PdO ^[31]. This result indicated that the Pd in the NP-Pd(II) was likely in the +2 oxidation state. In addition, the binding energy of this Pd 3d_5/2 peak was still 1.6 eV lower than that of the precursor PdCl₄^{2− [32]}, suggesting that the Pd in the catalyst was no longer coordinated to four Cl⁻ ligands.

Fig. 2. XRD (a) and XPS (b) patterns of NP-Pd (II).

The FT-IR spectrum of the NP-Pd(II) was acquired to determine the coordination structure of the Pd(II). The IR spectrum in Fig. 3(a) clearly shows two prominent bands at 1665 and 1605 cm⁻¹, attributed to the stretching vibrations of conjugated C=C bonds (ν(C=C)). A peak at 3016 cm⁻¹, corresponding to the stretching absorption of =C−H (ν(=C−H)), was also observed (Fig. S7). No absorption bands corresponding to the triple bond stretching mode (ν(C≡C)) of free acetylene (ca. 2100 cm⁻¹) or to coordinated Pd (ca. 1800 cm⁻¹) were present in the spectrum ^[33]. These results suggest that the polymerization of acetylene into conjugated alkenes took place in the presence of Pd(II), as has been well documented in the literature ^[34]. The formation of polyacetylene was also confirmed by the presence of IR absorption peaks in the region of 700 to 950 cm⁻¹, corresponding to the out-of-plane bending vibrations of =CH (δ(=CH)). Typically, the stretching mode of C=C generates medium or even weak peaks. In our case, however, the ν(C=C) peak appeared to exhibit the strongest IR absorption. This can be explained by the polarization of double bonds, likely caused by their coordination with Pd, and the conjugation of the C=C bonds in polyacetylene (vide infra). The polymerization of acetylene was further supported by solid-state ¹H NMR analysis of the NP-Pd(II), since the resulting spectrum showed a broad peak with a chemical shift centered at 5 ppm (Fig. S8), characteristic of the ¹H chemical shift of alkenes ^[35].

The coordination environment of the Pd was significantly changed with the formation of the NP-Pd(II), and so the UV-Vis absorption spectrum of the catalyst was different from that of the PdCl₄²⁻ precursor (Fig. S9). The aqueous PdCl₄²⁻ solution generated an absorption band at 218 nm with a shoulder at 243 nm, corresponding to the ligand-to-metal charge transfer absorptions of [PdCl₃(H₂O)]⁻ and [PdCl₂(H₂O)₂] ^[36]. In comparison, the NP-Pd(II) spectra contained only one prominent peak at 289 nm, clearly indicating a variation in the Pd coordination environment. The high background intensity in the NP-Pd(II) spectrum was ascribed to light scattering induced by NP-Pd(II) colloids. To further determine the Pd coordination in the NP-Pd(II), Raman spectroscopy was also employed. As can be seen from Fig. 3(b), the Raman spectrum of NP-Pd(II) displays two intense peaks, at 1260 and 423 cm⁻¹. The band at 423 cm⁻¹ is assigned to the Pd-C=C stretching mode, while the peak at 1260 cm⁻¹ is attributed to the strongly coupled C-C and C=C stretching modes ^{[37, 38]}. The weak peak at 274 cm⁻¹ is assigned to the Pd-Cl stretching mode ^[39]. These results indicate that the Pd atoms in the NP-Pd(II) were coordinated to both Cl and C=C moieties. The coordination of C=C to Pd is also supported by the presence of two weak bands at 1523 and 1428 cm⁻¹ in NP-Pd(II), corresponding to the stretching vibration of C=C coordinated with Pd. In FT-IR spectra, C=C bonds coordinated with Pd are expected to absorb at lower wavenumbers as compared to non-coordinated bonds ^[31]. Moreover, the solid state ¹³C NMR spectrum (Fig. S8) also confirmed the coordination of C=C bonds with Pd in the NP-Pd(II) based on the presence of a peak at a chemical shift of 80 ppm ^{[40, 41]}.

Fig. 3. FT-IR (a) and Raman (b) spectra of NP-Pd (II).

Since the first such report in 1984, there has been significant interest in the study of polymeric complexes obtained by reacting a Pd(II) precursor with various acetylene derivatives [34,42−44]. However, the detailed structures of these polymeric Pd complexes have rarely been studied by EXAFS ^[45]. To gain a deeper understanding of the local structure of the Pd in this material, fluorescence-mode XAS was also performed on the NP-Pd(II) at the Pd K-edge (24.35 keV) (Fig. S5). The radial distribution function obtained by Fourier transformation of the k³-weighted EXAFS data for NP-Pd(II) is shown in Fig. 4(a), together with results for Pd foil and Na₂PdCl₄ for comparison purposes. The EXAFS peak centered at approximately 2 Å can be attributed to the Pd-Cl coordination shell. A peak at the same position was also observed in the case of the Na₂PdCl₄ reference, in which Pd is coordinated only to Cl⁻. The shoulder at approximately 1.5 Å indicates the presence of very short Pd-ligand bonds, identified as a Pd-C coordination shell by the EXAFS fitting discussed below. The absence of Pd-Pd scattering further excludes the presence of metallic Pd in the NP-Pd(II). The XAS results also ruled out the presence of PdO (Fig. S10). Therefore, two coordination shells, Pd-C and Pd-Cl, were introduced to fit the EXAFS spectrum as a means of obtaining more precise coordination information for the Pd in the NP-Pd(II) (Fig. 4(b)). As shown in Table 1, quantitative analysis of the EXAFS spectra demonstrated that the Pd atoms had average coordination numbers of 2.2 and 2.4 for the Pd-C and Pd-Cl scattering paths, respectively (see also Supporting Information). Attempts to introduce an extra Pd-Pd coordination shell led to unreasonable fitting results.

Fig. 4. (a) EXAFS spectra of NP-Pd (II),Na2PdCl4,Pd foil;(b) curve fitting for the NP-Pd (II) data.

Table 1 Fitting results for the NP-Pd(II) EXAFS data.a Scattering path b C. N. c /atoms Bong length/Å Debye–Waller factor/Å2 E0-shift/eV Pd-C 2.2 ± 0.5 2.09 ±0.026 0.007 ±0.002 4 ± 2 Pd-Cl 2.4 ± 0.5 2.38±0.015 0.007 ±0.002 4 ± 2 a To reduce the number of free-running parameters,Debye-Waller factors and E0-shifts were assumed to be equal for both paths.b Scattering path of the photoelectron.c Coordination number.

Table 1 Fitting results for the NP-Pd(II) EXAFS data.a

Based on a combination of the spectroscopic results and EXAFS fitting data, we propose two possible models for the local structure of Pd in this material (Fig. S11). In the first model, Pd atoms are present as dimers bridged by two Cl⁻. In addition to one bridging Cl⁻, each Pd atom is also coordinated to one C=C unit and a terminal Cl⁻, giving coordination numbers of 2 and 3 for Pd-C and Pd-Cl, respectively. In the second model, one Pd atom is coordinated with two double bonds and two terminal Cl⁻ ligands, giving coordination numbers of 4 and 2 to Pd-C and Pd-Cl, respectively. It should be noted that both of these structures are common in acetylene complexes of Pd ^[46]. Reacting olefins with appropriate Pd precursors has also been shown to yield similar products ^[44]. The co-presence of both coordination patterns in the NP-Pd(II) was also indicated by the results of FT-IR spectroscopy. Absorption peaks corresponding to both coordinated (1523 and 1428 cm⁻¹) and free C=C bonds (1665 and 1605 cm⁻¹) were observed in the FT-IR spectrum of NP-Pd(II).

Based on the above analyses, the NP-Pd(II) comprised aggregated amorphous polyacetylene-supported Pd(II) nanoparticles, in which the Pd(II) atoms were atomically dispersed in the polyacetylene matrix. These structural features should make the NP-Pd(II) a desirable catalyst, and so Suzuki cross coupling reactions were performed using this material so as to evaluate its catalytic activity and recyclability. To fairly establish an activity ranking of NP-Pd(II) among different Pd catalysts, a model reaction employing iodobenzene and phenylboronic acid was first conducted in a mixed solvent of ethanol and deionized H₂O. Various other Pd catalysts, including Pd(OAc)₂, (CH₃CN)₂Cl₂Pd (denoted as CNCl-Pd), (Ph₃P)₄Pd and Pd/C (Aldrich, 10 wt% Pd on activated carbon) were chosen for comparison. As shown in Fig. 5(a), NP-Pd(II) exhibited catalytic activity comparable to that of each of the homogeneous catalysts (CNCl-Pd, (Ph₃P)₄Pd and Pd(OAc)₂) and remarkably better than that of Pd/C. After 0.5 h reactions, the iodobenzene conversions were 79%, 82%, 76%, 56% and 24% for NP-Pd(II), CNCl-Pd, Pd(OAc)₂, (Ph₃P)₄Pd and Pd/C, respectively. Thus, although NP-Pd(II) is essentially a heterogeneous catalyst, its activity was equal to or even better than that of homogeneous catalysts. Such high activity indicates that the Pd atoms in the NP-Pd(II) were highly accessible and that the C=C ligands derived from acetylene were effective for Suzuki coupling. In fact, the positive effect of C=C ligands on Pd-catalyzed cross-coupling has been recently summarized ^[47].

Fig. 5. Performance of NP-Pd (II) during the Suzuki coupling of iodobenzene with phenylboronic acid inmixed solvents (a,b) and H2O (c,d).Reaction conditions:(a)0.1 mol% Pd,10 mL ethanol,5 mL H2O,1 mmol iodobenzene,1.5 mmol phenylboronic acid,2 mmol Na2CO3,30℃.Reaction conditions changed relative to (a):(b)0.002 mol% Pd;(c) solvent:15 mL H2O;(d) solvent:15 mL H2O,60℃,1 h.

To illustrate the advantages of NP-Pd(II), the durability of the NP-Pd(II) catalyst was also compared with that of Pd/C and the most active homogeneous catalyst, CNCl-Pd, by reducing the Pd loading to 0.002 mol%. As shown in Fig. 5(b), after 2 h reactions, the NP-Pd(II) and CNCl-Pd gave iodobenzene conversion of 60 and 63%, respectively, while the Pd/C allowed only an 8% conversion. When the reaction time was extended, the CNCl-Pd did not produce appreciably higher conversions; after 12 h, the conversion over CNCl-Pd was only increased to 65%, indicating deactivation of the catalyst. In comparison, the NP-Pd(II) exhibited distinctly better performance, giving a conversion of 90% after 12 h. In other trials, a reaction temperature of 80 °C was found to generate almost complete conversion within 1 h. As well, further lowering the Pd loading to 0.00025 mol% gave 52% conversion within 1 h at 80 °C, corresponding to a TOF as high as 208000. This value is much greater than the highest value achieved using heterogeneous catalysts ^[48] and also shows a much more effective reaction promotion than that of many homogeneous catalysts ^[8]. After 6 h, 98% of the iodobenzene was converted, equivalent to a turnover number of 390000 (Fig. S12). Moreover, unlike many phosphine-based Pd complexes, the NP-Pd(II) was relatively stable. After storage for 6 months in open air at room temperature, no changes were observed in the FT-IR spectrum of the material nor did it exhibit any decrease in catalytic activity (Fig. S7 and S13).

Table 2 Suzuki coupling reactions with NP-Pd(II) in H2O. Entry Catalyst Ar-X T/°C t/h Yield a/% 1 2 3 4b 5 6 7 8 9 10 11 12 13 14 15 16 17 18c 19 20 21 22 NP-Pd(II) Pd/C CNCl-Pd CNCl-Pd NP-Pd(II) CNCl-Pd Pd/C NP-Pd(II) NP-Pd(II) NP-Pd(II) NP-Pd(II) NP-Pd(II) NP-Pd(II) NP-Pd(II) NP-Pd(II) NP-Pd(II) NP-Pd(II) NP-Pd(II) NP-Pd(II) NP-Pd(II) NP-Pd(II) NP-Pd(II) 70 70 70 70 40 40 40 60 60 70 80 70 70 40 70 80 80 80 60 70 70 70 3 3 3 3 1 1 1 1 1 3 8 3 8 1 3 6 8 10 8 3 3 10 97 64 56 94 98 94 64 99 99 97 82 99 99 94 98 96 51 91 95 98 99 75 Reaction conditions:0.1 mol% Pd,15 mL H2O,1 mmol benzene halide,1.5 mmol phenylboronic acid,2 mmol Na2CO3.R=-H,-Cl,-CH3 and-CH=CH2 for entries 1-19,20,21 and 22.a GC yield.b 10 mol% TBAB added.c 0.5 mol% Pd loading.

Table 2 Suzuki coupling reactions with NP-Pd(II) in H2O.

H₂O is an abundant, environmentally-friendly solvent and so there has been much research regarding Suzuki cross coupling reactions in aqueous systems [49−54]. To further demonstrate the advantages of NP-Pd(II), its catalytic performance in Suzuki coupling in water was investigated. CNCl-Pd has recently been used as a catalyst in aqueous Suzuki coupling ^[53], but the present work revealed that CNCl-Pd is less active than the NP-Pd(II) under such conditions. Trials showed that this material yielded only 20% conversion of iodobenzene within 1 h, while also clearly generating Pd black during the reaction. In comparison, under the same conditions, conversions of 63% and 41% were achieved when using the NP-Pd(II) and Pd/C, respectively (Fig. 5(c)). A similar trend was also observed for the coupling of bromobenzene with phenylboronic acid. While a biphenyl yield of 97% was achieved with the NP-Pd(II) in 3 h at 70 °C, the yields were 56% and 64% with the CNCl-Pd and Pd/C, respectively, under the same conditions (Table 2, entries 1-3). These results indicate that the NP-Pd(II) is a more effective catalyst in the aqueous phase. Furthermore, the NP-Pd(II) catalyst exhibited much better stability and thus recyclability in aqueous Suzuki coupling reactions. In these studies, products from each reaction were extracted using diethyl ether and the catalysts were re-used in a subsequent reaction. As clearly shown in Fig. 5(d), after five cycles, the NP-Pd(II) continued to catalyze the coupling reaction, with a conversion of 82%. In comparison, after five replicate reaction cycles, the Pd/C and CNCl-Pd gave conversions of only 28% and 17%, respectively. It is worth repeating that, due to its tendency toward aggregation, the NP-Pd(II) is easily separated from reaction solutions by centrifugation or simple settlement, making it highly applicable as a recyclable catalyst.

The NP-Pd(II) deactivation mechanism was also investigated. FT-IR spectra and elemental analysis did not indicate any obvious differences between the fresh and recycled catalysts (Fig. S7). However, high-resolution TEM (HRTEM) assessments found metallic Pd nanoparticles in the used catalyst (Fig. S14), a finding that was verified by XRD and XPS data (Fig. S15 and Fig. S16). The mechanism by which these nanoparticles were formed was investigated using XAS analyses of NP-Pd(II) samples subjected to different treatments. Combining the NP-Pd(II) with only Na₂CO₃, iodobenzene or phenylboronic acid in H₂O generated no changes in its EXAFS spectrum, indicating that the formation of metallic Pd was not induced by any individual component of the reaction mixture. However, following the reaction, the Pd-Pd coordination shell was clearly observed in the spectrum of used NP-Pd(II) (Fig. S17), demonstrating that the metallic Pd nanoparticles were formed during catalysis. Accordingly, we ascribe the gradual loss of activity of the NP-Pd(II) to the formation of less active Pd nanoparticles.

One of the most significant challenges associated with using aqueous Suzuki coupling is that the majority of halogenated aromatics are hydrophobic. As such, the lack of effective contact between the hydrophilic CNCl-Pd and the substrate was likely the main reason for its poor performance in H₂O. Accordingly, 10 mol% tetrabutylammonium bromide, a commonly used phase-transfer reagent for aqueous Suzuki coupling ^[55], was added to enhance the activity of the CNCl-Pd, and a significant improvement was obtained, increasing the yield from 56% to 94% (Table 2, entry 4). When the relatively hydrophilic substrate 4-bromophenol was used, high yields were achieved with both the NP-Pd(II) and CNCl-Pd (Entries 5 and 6). The superior activity of the NP-Pd(II) compared to CNCl-Pd was therefore attributed to its polyacetylene ligands and unique morphology, both of which allow more effective contact between the catalyst and hydrophobic substrates in H₂O.

To explore the scope of substrates to which the NP-Pd(II) is applicable, a wide range of aqueous Suzuki reactions was investigated. Using the NP-Pd(II), all reactions employing various aryl bromide or iodide substrates proceeded smoothly and gave excellent yields (Table 2, entries 8-16). It should be noted that the coupling of 2-bromotoluene required harsher conditions, indicating a possible steric effect in the reaction (Table 2, entry 11). In Suzuki coupling, chlorobenzenes are a challenging class of substrates to convert. However, employing NP-Pd(II) with 0.5 mol% Pd as the catalyst, a yield of 91% was obtained (Table 2, entries 17 and 18). Moreover, a sulfur-bearing substrate, 2-bromothiophene, was efficiently converted at 60 °C, with a yield of 95% achieved in 8 h (Table 2, entry 19). Phenylboronic acids with different substituents were also investigated; 4-Cl and 4-CH₃ phenylboronic acids were converted equally effectively, while 4-vinylphenylboronic acid was less active (Table 2, entries 20-22).

A facile and readily scalable synthetic strategy has been developed to yield an efficient nanoparticulate polyacetylene-supported Pd(II) catalyst for Suzuki coupling. The catalyst was obtained by simply treating an aqueous solution of PdCl₄²⁻ with acetylene under ambient conditions, and comprised micron-sized aggregates of fine nanoparticles of the Pd(II)-polyacetylene complex. Systematic structural characterization revealed that Pd(II) atoms in the catalyst were homogenously distributed in the polyacetylene matrix and coordinated with C=C and Cl ligands. The homogenous catalyst-like local structure and the heterogeneous catalyst-like morphology render the as-prepared catalyst both highly efficient and recyclable when applied to the aqueous Suzuki coupling of a wide range of aromatic substrates. In principle, almost all alkynes will readily react with Pd(II) to give polymeric complexes ^[34], thus the method reported herein is expected to be extendable to alkynes other than acetylene. By varying the acetylene substituent groups, it should be possible to synthesize more advanced Pd catalysts.

We thank Dr. Jyh-Fu Lee and Prof. Soofin Cheng for helpful discussions, and also the technical support from beamline 01C1 at National Synchrotron Radiation Research Center (Hsinchu). P.Z. also acknowledges the funding from NSERC Canada.