Bulky Di(1-adamantyl)phosphinous Acid-Ligated Pd(II) Precatalysts for Suzuki Reactions of Unreactive Aryl Chlorides.

Di(1-adamantyl)phosphine oxide (SPO-Ad: Ad2P(V)(=O)H), a stable tautomer of di(1-adamantyl)phosphinous acid (PA-Ad: Ad2P(III)-OH), was employed to synthesize two new PA-Ad-coordinated complexes, POPd-Ad and POPd2-Ad. POPd-Ad was easily transformed from POPd2-Ad in acetonitrile, and the [M - H]- ion of the deprotonated POPd-Ad was observed in the electrospray ionization-mass spectrum of POPd2-Ad. Both complexes are effective precatalysts for the Suzuki reaction of aryl chlorides. The reduction of Pd(II) in POPd-Ad and POPd2-Ad by arylboronic acid was examined, and the ideal Pd-to-PA ratio in the Suzuki reaction was found to be 1:1. The effect of temperature on the catalytic yields was studied to examine the possible ligation state of the active species and the dimer-to-monomer process of POPd2-Ad. Mononuclear and mono-ligated Pd species was assumed to be catalytically active. The electronic and steric effects of PA-Ad were slightly better than those reported for PA-tBu ( t Bu2P(III)-OH). Density functional theory calculations were performed to evaluate the formation of mono-ligated and mononuclear Pd species from POPd-Ad and POPd2-Ad. Furthermore, the reaction time and catalyst loading could be reduced for the reported POPd1-tBu precatalyst using the optimized reaction conditions for POPd-Ad. The complexes synthesized in this extensive study will complement the existing SPO-coordinated POPd series of precatalysts.

Introduction

Transition-metal (TM)-catalyzed cross-coupling reactions have been extensively researched because of their broad application areas such as in the synthesis of natural products[1] and in materials science.[2] Moreover, TM-catalyzed decarbonylative coupling of acyl fluorides[3] and N-acyl-amides,[4] as well as TM-catalyzed acylative[5] cross-coupling reactions have been extensively studied.[6] The successful advancements in these reactions are highly correlated to the continual progress of TM catalysts.[7] Auxiliary ligands, in particular, play a central role in the development of new TM catalysts.[8] For instance, phosphines,[9] N-heterocyclic carbenes (NHC),[10] and carbodicarbenes,[11] have been proven effective ligands for Pd-catalyzed cross-coupling reactions.

The phosphinous acid ligand (PA-R: R2P(III)-OH, with -R being the substitutions on PA) is a promising analogue of tri-substituted phosphines (PR3) (Scheme ).[12] PA-R can be easily tautomerized from the secondary phosphine oxide (SPO-R: R2P(V)(=O)H) in the presence of TM complexes (Scheme ).[13] Li et al. previously showed that di-tert-butyl phosphine oxide (SPO-tBu: Bu2P(=O)H) and its related Pd(II) complexes are effective preligands and precatalysts for various Pd(II)-catalyzed cross-coupling reactions.[14] Considering the stoichiometry and the bulkiness of substituents on SPO-R, three types of Pd(II) complexes can be obtained (Scheme ).[14b,14c] According to the literature as well as from reported experiments, SPO-R equipped with bulky substituents tends to form POPd-R and POPd2-R,[14a] while SPO-R functionalized with less bulky groups yields POPd-R and POPd1-R.[15] Similar observations were reported when sterically different SPO-R preligands were reacted with PtCl2(COD).[13a] Notably, PA-R has rich coordination modes, as reported in the literature.[16]

Scheme 1

Tautomerization of SPO-R to PA-R and Syntheses of Prototype PA-Coordinated Pd(II) Complexes

The steric effect of the two substituents on the coordinated PA-R has a noticeable impact on the catalytic capability of the associated Pd(II) precatalysts. For example, using aryl chloride as a substrate in the Suzuki reaction, PA-Ph-coordinated POPd1-Ph was reported as an inefficient precatalyst,[17] while the PA-tBu-coordinated POPd1-tBu was demonstrated to be a good precatalyst.[14c] Employing large sterically hindered SPO-R as a preligand, could therefore be considered a promising route to improve catalytic capability.[18] This was shown through the successful employment of secondary phosphine oxides with bulky substituents as preligands in several TM-catalyzed reactions.[19]

Sterically congested PA-Ad coordinated Complex 1 was demonstrated to be an effective precatalyst for the C–H bond arylation of oxazo(li)nes (Figure a).[20] Moreover, PA-diam coordinated Complex 2 was reported as a good precatalyst in the Kumada-Corriu cross-coupling reaction of 2-pyridyl Grignard reagents (Figure b).[19g] In addition, both complexes are suitable precatalysts for heteroarylation via the C–H activation step.[19i] Ackermann demonstrated that PA-Ad is an efficient ligand in Ru-catalyzed C–H bond functionalization,[21] Ni-catalyzed Kumada cross-coupling reactions,[22] and Pd-catalyzed intramolecular α-arylation reactions.[19f] Crystal structures for these two precatalysts were resolved.[19g,19i] Based on the literature, the SPO-Ad preligand has not been used to synthesize the POPd series of complexes. This gap in research might be due to the multiple steps required to synthesize SPO-Ad or the cost associated with SPO-Ad and its precursor di-1-adamantylphosphine chloride.

Figure 1

(a) PA-Ad and (b) PA-diam-ligated Pd(II) complexes.

The percent buried volume (%Vbur) parameter[23,24] has been employed to evaluate the steric effect associated with PA-R.[25a,18,25b] Examples where this parameter were successfully employed in the evaluation of the steric effect include the Au ← P(OH)R2 series of complexes where %Vbur parameters of 27.3% for P(OH)Ph2 and 32.6% for P(OH)Bu2 were reported.[25a] The %Vbur has also been employed in the evaluation of steric effects associated with PR3 and NHC ligands.[26] For Pd ← PR3 complexes, the %Vbur parameters for PAd3 and PBu3 were estimated to be 40.5 and 40.0%, respectively.[27] Therefore, it was expected that Ad2P-OH and Bu2P-OH would have comparable steric effects relative to the TM center.

To expand upon the POPd series of precatalysts as reported in the literature, it would therefore be reasonable to synthesize a POPd series of complexes employing a SPO-Ad preligand. Herein, we report the syntheses and crystal structures of such SPO-Ad-coordinated POPd-Ad and POPd2-Ad complexes. POPd2-Ad can be quickly converted to POPd-Ad in acetonitrile at room temperature. These two complexes are shown to be efficient precatalysts for the Suzuki reaction of aryl chlorides. The reduction of Pd(II) in POPd-Ad and POPd2-Ad, the Pd-to-ligand ratio in the catalytic cycle, the ligation state of potential catalytic species, and the possible reaction mechanism are suggested. Based on experimental and density functional theory (DFT) investigations, the formation of monoligated and mononuclear Pd(II) intermediates will be proposed.

Results and Discussion

Syntheses and Characterization of POPd-Ad and POPd2-Ad

Precatalyst POPd-Ad was synthesized through the reaction of PdCl2 and 2 equiv of SPO-Ad in a nitrogen-gas-filled Schlenk tube at 60 °C for 18 h (Scheme a). Because of the large steric hindrance associated with SPO-Ad, POPd1-Ad was not produced. Upon decreasing the amount of SPO-Ad to one equivalent and increasing the temperature to 110 °C for 15 h, POPd2-Ad was formed, which precipitated as the main product (Scheme b). Following workup and recrystallization processes, the yellow and orange solids produced were characterized as POPd-Ad and POPd2-Ad, respectively. The 31P NMR chemical shifts for POPd-Ad and POPd2-Ad were observed at 113 and 137 ppm, respectively. Additional supporting evidence include the DEPT-90, DEPT-135, and the full one-dimensional (1D) 13C NMR spectra for POPd-Ad and POPd2-Ad, as shown in Figure S1.

Scheme 2

Syntheses of POPd-Ad and POPd2-Ad

Employing a slow evaporation method under ambient conditions, crystals of POPd-Ad (Figure ) and POPd2-Ad (Table S1) were grown in CH2Cl2 for X-ray crystal structure analysis. The molecular structure of POPd-Ad indicates that Pd–P and Pd–Cl bond lengths are 2.3483(5) and 2.3082(6) Å, respectively. The two Pd–P bonds in POPd-Ad are longer than the average bond distance (2.2498 Å) associated with 50 Pd(II)–P(OH)R2 bonds[28] and are among the longest Pd(II) ← P(OH)R2 dative bond distances. This indicated a large steric repulsion between the two adjacent adamantyl substituents on the two Ad2P-OH ligands. In addition, the two different bond angles of Cl–Pd–P1 are 91.90(2) and 88.09(2)°, while those of P–Pd–P and Cl–Pd–Cl are exactly 180°. The distance between P–OH···Cl indicates the presence of two intramolecular hydrogen bonds (2.272 Å). It could therefore be concluded that the POPd-Ad with a d8 Pd(II) center exhibited a typical square-planar geometry. Although the formation of POPd2-Ad could be confirmed through characterization, the resolution of its X-ray structure was unsatisfactory, with an associated high R-value of 0.2081. Subsequent attempts to grow suitable POPd2-Ad crystals for confirmatory X-ray structure analysis were unsuccessful (Table S1). However, the X-ray structure of POPd2-Ad shows that the Pd2Cl2 core is butterfly-shaped, while each Pd center possesses a typical square-planar geometry.

Figure 2

Molecular structure of POPd-Ad. Some hydrogen atoms are omitted for clarity.

Although POPd2-Ad spontaneously dissociated to form POPd-Ad in acetonitrile (Method (1) in Scheme : 42%), this phenomenon was not observed in toluene, tetrahydrofuran (THF), or 1,4-dioxane (Table S2). Reacting POPd2-Ad with 2 equiv of SPO-Ad in acetonitrile for 5 min at room temperature resulted in a high yield of POPd-Ad (Method (2) in Scheme ; 90%). During the evaluation of electrospray ionization (ESI)-mass spectra of these products (Figures and S2), [M – H]− ions identified at m/z 813.33 and m/z 990.51 could be attributed to POPd-Ad and POPd2-Ad, respectively. In addition, [M – H]− ions at m/z 813.33 associated with POPd-Ad were also detected in the mass spectrum of POPd2-Ad. Several additional species detected in the mass spectrum of POPd2-Ad are presented in Figure . A comparison between the theoretical and observed isotope patterns was subsequently performed to validate the predicted structures. Here, ion species B, C, E, G, H, and I could be generated from the [M – H]− ion of POPd2-Ad, through dimer-to-monomer dissociation (B, C) or the dissociation of PA-Ad (E) or Cl– (G). Species H could be formed from the association of PdCl2 with G (Figure b), while species I could be generated from the recombination of D and G (Figure b). Attempts to predict the minor anionic species observed at m/z 631.08 were not successful. The observed [M – H]− ion of POPd-Ad was attributed to the combination of species B and F (PA-Ad). The B (PdCl2 ← P(O–)Ad2), C, and I were the three most abundant ionic species (Figure S2a).

Figure 3

Negatively charged species (a) B, C, E, G, and (b) H, I detected using the ESI mass spectrometer compared to isotope patterns of fragments from POPd2-Ad.

Assessment of the Electronic and Steric Effects of PA Ligands

Previously,[29] R2(HO)P → Ni(CO)3 was used to evaluate the electron-donating ability (v∼CO, A1 stretching mode) of PA-R. For nickel model complexes, the reported extrapolated Tolman electronic parameters (TEP, v∼CO*0.9540) for Bu2(HO)P and Ad2(HO)P were 2064 and 2061 cm–1, respectively. This result indicates that Ad2(HO)P is more efficient in donating electron density to the TM center as compared to Bu2(HO)P. Extrapolated TEPs of 2010 and 2009 cm–1 were also reported for Bu2P-O– and Ad2P-O– sets, respectively.[29] Because salt is present in many cross-coupling reactions, R2(HO)P may have deprotonated and interacted with cations in catalytic reactions. Based on this assumption and the theory employed by Martin et al.,[29] if the cation (K+) was considered in the R2(K+O–)P → Ni(CO)3 (R = Bu and Ad) model, the extrapolated TEPs were calculated to be 2035 and 2032 cm–1. Overall, the results indicated that (deprotonated) PA-Ad is a better electron donor than PA-tBu.

The steric effect of the different PA-R ligands substituted was again evaluated by the percent buried volume (%Vbur). Using the crystal structures of POPd-tBu[30,25b] and POPd-Ad, the %Vbur values were estimated to be 29.1% and 30.2%, respectively. As for POPd2-tBu[14a] and POPd2-Ad, the %Vbur values were estimated to be 32.1 and 32.5%, respectively. The steric effect because of the PA-Ad ligand was slightly larger than that due to PA-tBu. In terms of steric and electronic descriptors, PA-Ad can therefore be expected to be a suitable ligand for palladium.

POPd-Ad and POPd2-Ad as Precatalysts for Suzuki Cross-Coupling Reactions

POPd-Ad was utilized to test its potential as a precatalyst for the Suzuki reaction of phenyl chloride (Scheme ). To identify the best combination of the solvent and base, initial experimental conditions included the use of 2 mol % of POPd-Ad, 0.5 mmol of aryl chloride, 0.55 mmol of 4-tolylboronic acid, and 3 equiv of base. The reaction was conducted in 2 mL of solvent at 65 °C for 12 h while evaluating the effect of different solvents.

Scheme 3

Initial Reaction Condition Employed To Screen the Best Combination of the Base and Solvent for the Model Suzuki Reaction

For the four solvents evaluated (toluene, THF, 1,4-dioxane, and acetonitrile) (Table and Table S3), toluene and acetonitrile were shown to not be suitable solvents for the Suzuki reaction, although six bases were examined for each (Tables and S3). For THF and 1,4-dioxane, the best conversion yields (up to 91%) were obtained with KOBu. To optimize the reaction time for the Suzuki reaction, further work was therefore focused on these two solvents. Here, the experiments indicated a marginal improvement in the reaction rate for 1,4-dioxane compared to that of THF over the first 3 h of the reaction (Table S4). Hence, 1,4-dioxane and KOBu were selected as optimum solvents when employing POPd-Ad as the precatalyst.

Table 1

Search for an Optimized Set of Base and Solvent for POPd-Ad-Catalyzed Suzuki Reaction of Phenyl Chloridea,b

	base	solvent
		toluene	THF	1,4-dioxane	ACN
1	KF	1	49	17	<1
2	K2CO3	2	11	32	<1
3	K3PO4	9	62	66	<1
4	KOtBu	44	91	91	ND
5	CsF	17	30	50	7
6	Cs2CO3	14	54	47	<1

Reaction conditions: Phenyl chloride (0.5 mmol) and 4-tolylboronic acid (0.55 mmol) were employed as substrates. Reactions were carried out at 65 °C for 12 h.

Conversions (%) were determined by GC–MS.

A low gas chromatography–mass spectrometry (GC–MS) conversion (7%) was obtained when phenyl chloride was catalyzed with acetonitrile and CsF at 65 °C for 12 h (Scheme ). However, when Ph-Br (62%) and Ph-OTf (32%) were used as the substrate while the same reaction conditions were employed, much higher isolated yields were obtained. This result agrees with previous results reported by Proutiere and Schoenebeck, indicating that the oxidative addition of the C–OTf bond to Pd(0) catalytic species in the presence of F– or ArB(OH)O– in polar acetonitrile was preferable for the use of nonpolar THF.[31] This C–Cl or C–OTf bond preference may be due to the formation of coordinating anionic species (ArB(OH)O– or F–) in polar solvents[31] and the related bis-ligated catalytic species.[32]

Finally, the Suzuki reactions of phenyl chloride and 1.1 equiv of 4-tolylboronic acid were carried out in 1,4-dioxane with KOBu at either 85 or 95 °C (Table S5). After 2 h, the GC conversion yields were 68% at 85 °C and 75% at 95 °C. As a final optimization step, the amount of 4-tolylboronic acid was increased to 1.5 equiv, and this amount was reacted at 95 °C for 0.5 h, resulting in a conversion of 99% (Table S6). The results further indicated that the reactions were completed in 0.5 h. The optimized reaction conditions for the POPd-Ad precatalyst are summarized in Scheme .

Scheme 4

Optimized Reaction Conditions for POPd-Ad and POPd2-Ad-Catalyzed Suzuki Reaction

Di-nuclear POPd1-tolyl is ligated by PA-Ph (Ph2(HO)P:), with two smaller phenyl groups,[33] while mononuclear POPd-tBu,[34] di-nuclear POPd1-tBu, and POPd2-tBu[14a,34] are associated with bulky Bu2P-OH (PA-tBu) ligands. The latter precatalysts have previously been demonstrated to be effective precatalysts for Suzuki reactions of aryl chlorides.[14c] However, to achieve good catalytic yields, long reaction times were required. In this study, these four complexes were synthesized based on previously reported methods and subsequently used as precatalysts employing the optimized conditions determined for POPd-Ad (Table ).

Table 2

Comparison of the Suzuki Reactions of Phenyl Chloride and 4-Tolylboronic Acid as Catalyzed Using Various Precatalystsa,b

entry	precatalyst	mol %	isolated yield (%)
1	POPd1-tolyl	1	15
2	POPd-tBu	2	72
3	POPd1-tBu	1	81
4	POPd2-tBu	1	70
5	POPd-Ad	2	99
6	POPd2-Ad	1	99

Phenyl chloride (0.5 mmol) and 4-tolylboronic acid (0.75 mmol) were used as substrates.

The optimized reaction conditions are shown in Scheme .

As expected, the three PA-tBu-coordinated Pd(II) complexes were found to be satisfactory precatalysts (entries 2–4), although POPd1-tolyl was not effective. By employing the optimized reaction conditions, the reaction time for the POPd1-tBu-catalyzed Suzuki reaction could further be significantly reduced from 12 h (previously reported) to only 0.5 h.[14a] In addition, both POPd-Ad and POPd2-Ad were more efficient precatalysts than POPd1-tBu and POPd2-tBu, as shown by the quantitative conversion of phenyl chloride to 4-phenyltoluene (entries 5 and 6). Bis-ligated POPd-Ad was shown to be a less efficient precatalyst than monoligated POPd2-Ad.

Ligation State of Active Species in the Suzuki Cross-Coupling Reaction

To determine the influence of the amount of the SPO-Ad preligand on the catalytic yield of the Suzuki reaction shown in Scheme , excess SPO-Ad was added to the POPd2-Ad catalyzed Suzuki reaction of phenyl chloride and 4-tolylboronic acid (Table ). Here, the use of excess preligands resulted in a reduction of the catalytic yield, with the yield associated with 4-phenyltoluene decreasing to only 5% when 10 mol % of SPO-Ad was added (entry 4: Pd:SPO-Ad = 1:5). These results indicated that the active species could be a mononuclear and monoligated Pd complex.

Table 3

Determination of the Optimal Pd-to-SPO-Ad Ratio for the POPd-Ad-Catalyzed Suzuki Reaction of Phenyl Chloride and 4-Tolylboronic acida

entry	POPd2-Ad (mol %)	SPO-Ad (mol %)	conversion yieldb (%)
1	1	0	95
2		2	80
3		4	27
4		10	5

Phenyl chloride (0.5 mmol), 4-tolylboronic acid (0.55 mmol), and 3 equiv of KOBu; reaction in 2 mL of 1,4-dioxane at 65 °C for 12 h.

Determined by GC with a barrier discharge ionization detector (GC-BID).

DFT Calculations and Temperature Effects on the Formation of Mono-Ligated Pd(II) Species

To gain insight into the differences between the Ad2P-OH and Bu2P-OH ligands in catalysis, DFT calculations were performed to determine the ligand dissociation free energies for POPd-R (Figure ) and the dimer-to-monomer process free energies for POPd2-R (Figure ). The most stable isomers identified were associated with species with two intramolecular hydrogen bonds between the two P–OH···Cl fragments, and these were consequently evaluated in further investigations (Figures S3–S5).

Figure 4

Gas-phase (solution) dissociation free energies for (a) POPd-Ad and (b) POPd-tBu (ωB97xD results, MWB28 pseudopotential and valence basis set for Pd, and 6-31G(d) for the other atoms. IEFPCM model was used to account for the solvation effect of 1,4-dioxane.).

Figure 5

Gas-phase (solution) dissociation free energies for (a) POPd2-Ad and (b) POPd2-tBu (ωB97xD results, MWB28 pseudopotential and valence basis set for Pd, and 6-31G(d) for the other atoms. The IEFPCM model was used to account for the solvation effect of 1,4-dioxane).

Here, the ωB97xD-optimized geometries of POPd-Ad (Figure a) and POPd-tBu (Figure b) well fit the structures determined by X-ray analysis. For POPd-tBu, the optimized (X-ray) bond lengths for the two Pd–P bonds are 2.370 Å (2.3450(3) Å), while they are 2.354 Å (2.3132(3) Å) for two Pd–Cl bonds.[30] Moreover, the two P–O bond lengths are 1.635 Å (1.6141(17) Å), with the two associated Cl–Pd–P angles being 88.3° (88.98(1)°). In the case of POPd-Ad, the optimized (X-ray) bond lengths for the two Pd–P bonds are 2.369 and 2.372 Å (2.3483(5) Å), and 2.356 and 2.358 Å (2.3082(6) Å) for the two Pd–Cl bonds.[30] Here, the two P–O bond lengths are 1.636 Å (1.6141(17) Å), and the two Cl–Pd–P angles are 88.3° (88.09(2)°).

Based on the X-ray structure reported for POPd2-tBu, it has a perfect planar Pd2Cl2 core.[14a] Calculations performed as part of this work did, however, indicate that this planar core is a transition state with an out-of-plane imaginary frequency of −7.97 cm–1 existing between two PdCl2(P(OH)Bu2) fragments. The observed planar Pd2Cl2 core could, therefore, have been a result of the solid-state packing force. The local minimum of POPd2-tBu is associated with a butterfly-shaped Pd2Cl2 core (Figure b). Both the X-ray structure and DFT-optimized geometry for POPd2-Ad confirmed this proposed geometry (Figure a and Table S1). The butterfly-shaped Pd2Cl2 core in POPd2-tBu and POPd2-Ad could be due to the steric hindrance between the two adjacent Bu2P-OH and Ad2P-OH ligands. The optimized (X-ray) Pd···Pd distances for POPd2-tBu and POPd2-Ad are 3.110 and 3.068 Å (3.046(3) Å), respectively, with Pd–P distances for POPd2-tBu and POPd2-Ad being 2.270/2.272 and 2.262 Å (2.235(6) and 2.236(8) Å), respectively. The two Pd–Cl distances associated with the connection of two R2(OH)P → Pd(II)Cl2 fragments for POPd2-tBu 2.467 and 2.468 Å, compared to the associated Pd-Cl distances of 2.475 Å (2.452(8) and 2.466(9) Å) for POPd2-Ad, respectively. Because of their different core geometries, the X-ray structure of POPd2-tBu was not compared with its optimized counterpart.

The bond dissociation free energies of PA-R for POPd-Ad and POPd-tBu in 1,4-dioxane were estimated to be 27.9 and 31.2 kcal/mol, respectively (Figure ), indicating that the dissociation of PA-Ad from the Pd(II) center to form monoligated and mononuclear Pd species would be easier than that for PA-tBu. In contrast, the two dimer-to-monomer dissociation free energies of POPd2-Ad and POPd2-tBu were estimated to be 22.2 and 18.6 kcal/mol, respectively (Figure ). As indicated by the interaction region indicator (IRI),[35] the smaller bond dissociation free energy of POPd-Ad, when compared to that of POPd-tBu, may originate from the stronger steric repulsion between two closely contacted adamantyl groups when attached to two different PA ligands (Figure S6a,b). Here, the two adamantyl units in the mononuclear POPd-Ad are closer to each other than in the di-nuclear POPd2-Ad. IRI analyses further indicated that the larger dimer-to-monomer free energy of POPd2-Ad, compared to that of POPd2-tBu, might be due to van der Waals interactions between the two adamantyl groups attached to two different PA ligands (Figure S6c,d).

Considering that the ligated R2(HO)P is a deprotonatable ligand, the dissociation free energies for the singly deprotonated POPd-R (R2(−O)P → PdCl2 ← P(OH)R2) and the dimer-to-mono process for the deprotonated POPd2-R ((R2(−O)P → PdCl)(μ-Cl)2 (PdCl←P(OH)R2)) in 1,4-dioxane were calculated (Figures S7 and S8). In Figure S7, the isomers with the best stability were also compared. For the singly deprotonated POPd-Ad (Figure S7a) and POPd-tBu (Figure S7b), the ligand dissociation free energies of PA-R in 1,4-dioxane were 9.5 and 11.7 kcal/mol, while the dissociation free energies associated with the breaking dimer for deprotonated POPd2-Ad and POPd2-tBu in 1,4-dioxane were 15.3 and 12.1 kcal/mol (Figure S8), respectively. The smaller free energies required for the dissociation of the singly deprotonated POPd-R and POPd2-R, when compared to that of their neutral counterparts, may be due to the stronger trans effect of the deprotonated R2(−O)P ligand. Notably, the resultant deprotonated mono-PA-ligated and mononuclear Cl2Pd(II) → P(O–)Ad2 species is the ionic species B detected in the ESI– mass spectrum of POPd2-Ad. In addition, the negative dissociation free energies calculated for the fully deprotonated POPd-R (ligand dissociation) and POPd2-R (dimer-to-monomer process) in 1.4-dioxane indicated that these were spontaneous reactions (Figures S11 and S12).

By employing phenyl chloride and 4-tolylboronic acid as substrates in the Suzuki reaction, the reaction temperature of 95 °C was lowered to study the ligation state of POPd-Ad and POPd2-Ad (Table S7). At 45 and 55 °C, the respective GC yields of the POPd-Ad-catalyzed Suzuki reaction were 3 and 54%, respectively. Compared to this, GC yields of the POPd2-Ad catalyzed reaction were 10 and 83% at 45 and 55 °C, respectively. This indicated that POPd2-Ad performed better as a precatalyst when compared to POPd-Ad. In addition, the dissociation of the PA-R ligand from POPd-R became unfavorable at lower temperatures, while the dissociated PA-R further reduced the catalytic yield. (Table ). These resultant species could be monoligated and mononuclear R2(−O)P → Pd(II)Cl2 for both precatalysts. It should further be considered that the ligand dissociation free energies for the singly deprotonated POPd-Ad and POPd2-tBu are 9.5 and 11.7 kcal/mol, respectively. Combined with the larger electronic and steric effects associated with PA-Ad when compared to PA-tBu, this supported the conclusion that POPd-Ad would be a better precatalyst than POPd-tBu (entries 2 and 5 in Table ).

Nonetheless, in 1,4-dioxane, the dimer-to-monomer free energy of the singly deprotonated PAPd2-Ad (15.3 kcal/mol) was found to be larger than the ligand dissociation free energy of the singly deprotonated POPd-Ad (9.5 kcal/mol). The energy required for the formation of R2(−O)P → Pd(II)Cl2 therefore appeared to be contradictory to the catalytic yields observed at 45 and 55 °C. Table further shows that the best Pd-to-PA ratio is 1:1, with yields being significantly affected by the redundant ligands, as seen in the low yield of the POPd-Ad-catalyzed Suzuki reaction at 45 and 55 °C. At 95 °C, the catalytic performances of POPd-Ad and POPd2-Ad were comparable (entries 5 and 6 in Table ). This would indicate that elevated temperatures prevent the re-coordination of extra Ad2(HO)P or Ad2(−O)P ligands in POPd-Ad-catalyzed Suzuki reactions.

Reduction of Pd(II) Precatalysts

Potential reduction pathways of Pd(II) in POPd-Ad and POPd2-Ad were also investigated (Scheme , results listed in Table ). Here, either POPd-Ad or POPd2-Ad was reacted in 1,4-dioxane with four equiv of 4-tolylbronic acid and 30 equiv of KOBu at 95 °C for 0.5 h. Three products were obtained from these reactions. First, the formation of 4,4′-dimethyl-1,1′-biphenyl may be associated with the reduction of Pd(II) by 4-tolylbronic acid in either POPd-Ad or POPd2-Ad.[36] The isolation of phosphine oxide further indicated that Pd(II) in the precatalyst was reduced through Ctolyl-P(=O)Ad2 bond coupling. This agrees with a reductive elimination reaction via C–P bond formation between the deprotonated SPO-R and the aryl group of the aryl-Pd(II)-P(O–)R2 species, as has been reported in the literature.[37] Finally, the dissociation of PA-Ad may be assumed, where SPO-Ad could be isolated, although the decomposition of Pd species cannot be excluded as a possible mechanism.

Scheme 5

Reaction between the POPd-Ad and POPd2-Ad with 4-Tolylboronic Acid in the Presence of 30 equiv of KOBu at 95 °C for 0.5 h

Table 4

Reaction between the POPd-Ad or POPd2-Ad with 4-Tolylboronic Acid in the Presence of 30 equiv of KOBua,b

	precatalyst	biphenyl (%)	phosphine oxide (%)	SPO-Ad (%)
1	POPd-Ad	14	ND	30
2	POPd2-Ad	73	ND	5

See Scheme for the detailed reaction condition.

Isolated yield.

After the reaction shown in Scheme was completed, 4,4′-dimethyl-1,1′-biphenyl was isolated. With no evidence of the formation of phosphine oxide (Tables and S8), boronic acid was assumed to be the reducing agent in the reduction of Pd(II) in POPd-Ad and POPd2-Ad. SPO-Ad could further be isolated from both POPd-Ad (30%) and POPd2-Ad (5%). These results imply that the dissociation of PA-Ad from bis-ligated and mononuclear POPd-Ad is crucial to form monoligated Pd(0) active species. Alternatively, this could also represent the decomposition of the reduced Pd(0) active species. However, it is believed that most SPO-Ad dissociated from POPd-Ad to form the monoligated and mononuclear species at 95 °C. This is supported by the fact that only 5% of SPO-Ad was recovered when the dimer-to-monomer process of POPd2-Ad was conducted under the same reaction conditions employed for POPd-Ad.

Yields of 4,4′-dimethyl-1,1′-biphenyl were 14 and 73% for POPd-Ad and POPd2-Ad, respectively. For POPd-Ad, the low yield could be attributed to the bis-ligated nature of the precatalyst, while the dissociated SPO-Ad could also recoordinate to the Pd(II) center to slow down the rate of the Pd(II) reduction of POPd-Ad. Regarding POPd2-Ad, the higher yield may be associated with the monoligated Pd(II) species formed after the dimer-to-monomer process of POPd2-Ad.

Application of POPd-Ad and POPd2-Ad to Other Aryl Chlorides

To examine the applicability and functional group tolerance of the substrates, POPd2-Ad was employed as a precatalyst in the Suzuki reaction (Figure ). The use of electron-deficient 4-(trifluoromethyl)phenylboronic acid resulted in a low yield (c: 31%).[38] Comparing the yields of products a–c, the use of 4-tolylboronic acid (b) was associated with the best result.[38] However, poor yields were obtained when a sterically congested arylboronic acid was selected as a substrate [e (99%) versus f (27%)]. Considering the yields of products a, d, e, g, h, and i, increasing the steric hindrance on the 2- and/or 6-position of aryl chlorides appears not to influence the yields significantly. This would imply that the oxidative addition of aryl chlorides to the Pd(0) active species proceeded smoothly. Data for g (81%), h (71%), and j (78%)-k (71%) revealed the same trend. Comparing the yields of i (74%) and l (24%), the low yield of l could be related to the synergy between steric and electronic effects of naphthalen group. Considering both the electronic and steric effects associated with the yields associated with arylboronic acids, the transmetalation may be assumed to be the rate-limiting step.[39]

Figure 6

POPd2-Ad-catalyzed Suzuki reactions of various aryl chlorides and boronic acids carried out employing the optimized reaction conditions. Isolated yields were reported. aThree equivalents of CsF were used as the base. The duration of these reactions was 1 h.

Good yields of products m (83%) and n (92%) were further observed where the electron-donating group was functionalized on the 4-position of phenyl chloride (Figure ). For heteroatom-substituted 2-chloropyridine, a yield of 99% was obtained (o). The use of 2-chloropyridine in POPd2-Ad catalyzed Suzuki reaction could therefore be proposed as an alternative route to overcome the problematic protodeboronation of 2-pyridyl boronic acid.[40] For 4-chlorobenzonitrile, the yield of 4-cyano-4′-methylbiphenyl was found to be 56% (p), while no conversion of 4-chloroacetophenone (q) was observed. This could be related to the use of –OBu. CsF was subsequently used to replace KOBu in the two reactions, where, after only 1 h, yields of p and q increased to 75 and 78%, respectively. No conversion was observed when benzyl chloride was used as the substrate.

Reaction Mechanism

Li et al. reported that Bu2(O–)P → Pd(0) ← P(O–)Bu2 or anionic Bu2(O–)P → Pd(0)Cl ← P(O–)Bu2 could precipitate when POPd1-tBu was reacted with a base.[14b] Here, it was, however, assumed that the transmetalation reaction between the Pd(II) species and aryl boronic acid could also occur after the di-nuclear POPd2-Ad dissociated into two mononuclear Cl2Pd(II) ← P(OH)Ad2 or deprotonated Cl2Pd(II) ← P(O–)Ad2 species (species B detected in the ESI– mass spectrum of POPd2-Ad).[41] In addition, for the mononuclear POPd-Ad, one bulky PA-Ad could initially dissociate from the Pd(II) center to form the same mononuclear Pd(II) species.

It was determined that when the base was reacted with POPd1-Ph ((Ph2(HO)P)2PdCl2), the deprotonation of R2(HO)P occurred to yield polymeric (Ph2(−O)P)2Pd(II).[15] Therefore, an anionic mononuclear Cl2Pd(II) ← P(O–)Ad2 was proposed when the base was reacted with POPd-Ad and POPd2-Ad. Subsequently, an anionic (4-tolyl)-ClPd(II) ← P(O–)Ad2 intermediate could be formed through a base-assisted transmetalation reaction with 4-tolylboronic acid. The reduction of Pd(II) would then proceed via a second base-assisted transmetalation reaction between boronic acid and the (4-tolyl)-ClPd(II)-P(=O)Ad2 intermediate to yield (4-tolyl)2-Pd(II)-P(=O)Ad2 species. Following the reductive elimination of 4,4′-dimethyl-1,1′-biphenyl, the anionic and zero-valent active species of Pd(0) ← P(O–)Ad2 would then be formed (Figure ), where the electron-rich anionic P(O–)Ad2 could facilitate the oxidative addition of unreactive aryl chlorides. Based on the results obtained in this study, the proposed reaction mechanism for the POPd-Ad- and POPd2-Ad-catalyzed Suzuki reaction of aryl chloride is shown in Figure .

Figure 7

Proposed Suzuki reaction mechanism as catalyzed by the POPd2-Ad. The graphical representation of the catalytic cycle was generated using the catalytic cycle generator.[42,43]

Conclusions

In this study, POPd-Ad and POPd2-Ad were synthesized for the first time. The di-nuclear POPd2-Ad could be easily converted into two mononuclear POPd-Ad monomers in acetonitrile. POPd-Ad could also be obtained from the reaction of POPd2-Ad and SPO-Ad. It was shown that both POPd-Ad and POPd2-Ad complexes were useful precatalysts for Suzuki cross-coupling reactions of aryl chlorides in only 0.5 h. The potential reduction pathway and the ligation state for POPd-Ad as well as POPd2-Ad were investigated, showing that the Pd(0) active species was reduced by arylboronic acid. Here, the monoligated Pd(0) ← P(O–)Ad2 was proposed to be the active species. DFT calculations and experimental data showed that the dissociation of the redundant PA-Ad ligand had a negative impact on the performance of the overall reaction when POPd-Ad was employed. Therefore, POPd2-Ad was considered a better precatalyst than POPd-Ad. A catalytic reaction mechanism was further proposed. Employing the optimized reaction conditions obtained here, the reaction time required for POPd-tBu-, POPd1-tBu-, and POPd2-tBu-catalyzed Suzuki reactions could be significantly reduced from 12 h to approximately 0.5 h. These PA-Ad-ligated Pd(II) complexes will complement current research on the POPd series of precatalysts.

Experimental Section

The experimental and computational methods, characterizations, NMR spectra, and optimized coordinates are thoroughly detailed in the Supporting Information.