A Dichotomy in Cross-Coupling Site Selectivity in a Dihalogenated Heteroarene: Influence of Mononuclear Pd, Pd Clusters, and Pd Nanoparticles—the Case for Exploiting Pd Catalyst Speciation

Site-selective dihalogenated heteroarene cross-coupling with organometallic reagents usually occurs at the halogen proximal to the heteroatom, enabled by intrinsic relative electrophilicity, particularly in strongly polarized systems. An archetypical example is the Suzuki–Miyaura cross-coupling (SMCC) of 2,4-dibromopyridine with organoboron species, which typically exhibit C2-arylation site-selectivity using mononuclear Pd (pre)catalysts. Given that Pd speciation, particularly aggregation, is known to lead to the formation of catalytically competent multinuclear Pdn species, the influence of these species on cross-coupling site-selectivity remains largely unknown. Herein, we disclose that multinuclear Pd species, in the form of Pd3-type clusters and nanoparticles, switch arylation site-selectivity from C2 to C4, in 2,4-dibromopyridine cross-couplings with both organoboronic acids (SMCC reactions) and Grignard reagents (Kumada-type reactions). The Pd/ligand ratio and the presence of suitable stabilizing salts were found to be critically important in switching the site-selectivity. More generally, this study provides experimental evidence that aggregated Pd catalyst species not only are catalytically competent but also alter reaction outcomes through changes in product selectivity.


Instrument Details and Methods for Compound Characterisation
NMR spectra were obtained in the solvent indicated in the text below, using a Bruker AVIIIHD 500 instrument (500  ). Chemical shifts (δ) are reported in parts per million (ppm) and were referenced to the residual non-deuterated solvent of the deuterated solvent used; CHCl3 : δ 1 H = 7.26 and 13 C = 77.16 (CDCl3), CD2Cl2 : 1 H = 5.31 (CDHCl2) and 13 C = 54.0, THF-d8 δ 1 H = 3.59 (OCH2CH2), 13 C = 67.57 OCH2CH2), 1 H = 1.73 (OCH2CH2) 13 C = 25.37 (OCH2CH2). Spectral data was typically collected at 298 K (25 o C). 31 P NMR spectral data were collected with proton decoupling, unless otherwise stated. 31 P NMR spectra were typically recorded using 128 scans and a spectral window of 300 ppm (δ 250 to -50 ppm). Chemical shifts for 31 P resonances were calibrated by externally referencing to an 85% H3PO4 in H2O (w/w). This was practically carried out by inserting a sealed, vacuum-dried capillary tube containing 85% H3PO4 in H2O (w/w) into an NMR tube containing the sample of interest, collecting a 31 P NMR spectrum and setting the H3PO4 resonance to 0 ppm. All 31 P and 13 C NMR spectra were obtained with 1 H decoupling. All NMR spectra were processed using MestReNova (MNova) software (using versions [12][13][14]. HRMS ESI-MS spectra were measured using a Bruker Daltronics micrOTOF MS, Agilent series 1200LC with electrospray ionisation (ESI) or on a Thermo LCQ using electrospray ionisation, with <5 ppm error recorded for all HRMS samples. LIFDI mass spectrometry was carried out using an JEOL AccuTOF GCx-plus instrument (JMS-T200GC), fitted with a probe produced by Linden CMS. The probe was equipped with 13 µm emitters on an AccuTOF. Alternatively, LIFDI-MS was carried out using a Waters GCT Premier MS Agilent 7890A GC instrument.
Mass to charge ratios (m/z) are reported in Daltons. High resolution mass spectra (HRMS) are reported with <5 ppm error (ESI and LIFDI). For clarity, LIFDI data are reported for 106 Pd, the most abundant natural isotope of Pd, which is part of 'exact mass' values.
Infrared spectra were obtained using a Bruker ALPHA-Platinum FTIR Spectrometer with a platinum-diamond ATR sampling module.
Melting points were recorded using a Stuart digital SMP3 melting point analysis machine.
Elemental analysis (Carbon, Hydrogen and Nitrogen {CHN} content) was carried out on an Exeter Analytical Inc.

CE-440 analyser.
For single crystal X-ray crystallographic analysis details, see Section 4. An oven-dried Schlenk tube (Flask 1) charged with Pd3(OAc)6 (6.7 mg, 0.03 mmol {3 mol%/Pd}) and triphenylphosphine (Table S1) was evacuated and backfilled with N2. THF (2.5 mL; dry, degassed) was added via a syringe and the immediately formed greenish-yellow suspension was stirred in an oil bath which was pre-heated to 40 o C. Another oven-dried Schlenk tube (Flask 2) was charged with 2,4-dibromopyridine (236.9 mg, 1.00 mmol, 1 equiv.) and p-fluorophenylboronic acid 2a (167.9 mg, 1.20 mmol) and subsequently evacuated and backfilled with N2. After 30 minutes of stirring, the contents of Flask 1 were transferred into Flask 2 via a cannula and the resulting mixture was stirred for 5 mins in order to thermally equilibrate at 40 o C. Aqueous Tetra-Nbutylammonium hydroxide (2.5 mL, 1.0 M, degassed) was then added (t0) to commence the reaction.
Samples (ca. 125 µL), taken at the specified times using a 1000 μL syringe were rapidly quenched by dissolution in CH2Cl2 solution before being filtered through a pipette fitted with a Celite® plug (ca. 1 cm depth) (to remove black particles). The filtrate was concentrated in vacuum to reveal a reddish-brown oil which was dissolved in CDCl3 (0.5 mL) for NMR sampling. Samples were analysed according to the procedure highlighted in Section 2.7.

Base and Additive Effects -SMCC Reactions (Table 1 -Main Paper) Scheme S2 Conditions and reagents for the investigation into base and additive effects on the model siteselective SMCC reaction at 1, catalysed by [Pd3(μ-Cl)(μ -PPh2)2(PPh3)3]Cl or Pd(OAc)2/1 or 2 PPh3.
An oven-dried Schlenk tube (Flask 1) charged with 2,4-dibromopyridine 1 (236.9 mg, 1.0 mmol), para-anisyl boronic acid 2b (182.4 mg, 1.2 mmol) Pd Cat. (3 mol% Pd loading; see Table S2 for specifics) was evacuated and backfilled with N2. Depending on the reaction, the salt additive was added at this point (See Table S3 for amounts used). The flask atmosphere was then evacuated and backfilled with N2 before THF (2.5 mL; dry, degassed) was added via a syringe and stirred for 5 minutes at 40 o C (pre-heated oil bath). After this time, aqueous base (2.5 mL, 1.0 M, degassed) was added (t0) to commence the reaction. The resulting reaction solution was hence stirred at 40 o C. Samples (ca. 125 µL), taken at the specified times using a 1000 μL syringe were rapidly quenched by S6 dissolution in CH2Cl2 solution before being filtered through a pipette fitted with a Celite® plug (ca. 1 cm depth) (to remove black particles). The filtrate was concentrated in vacuum to reveal a reddish-brown oil which was dissolved in CDCl3 (0.5 mL) for NMR sampling. See Section 2.7, for details of sample analysis.  Table 4) were added. The flask was sealed (Suba-seal®), THF (2.5 mL; dry, degassed) was added using a syringe and the resulting mixture was magnetically stirred for 5 minutes, in a pre-heated oil bath. After this time, all solids were seen to have dissolved into a clear S7 solution and the temperature of the reaction was measured at 40 ± 0.5 °C using an internal thermocouple. The cross-coupling reaction was then initiated by addition of tetra-n-butylammonium hydroxide solution (aq. 2.5 mL, 1.0 M→0.5 M, 2.5 equiv.) which was added by a rapid injection using a syringe over the Suba-seal®. The reaction was stirred, and samples were taken at the 0, 2.5, 5, 7.5, 10, 15, 20, 25, 20, 40, 50, 60, 80 and 100 minutes.
Samples (ca. 100 µL), taken at the specified times via a 1000 μL syringe were rapidly quenched by dissolution in CH2Cl2 before being filtered through a pipette fitted with a Celite® plug (ca. 1 cm depth) (to get rid of any particulate Pd). Each filtrate was concentrated in vacuum to reveal a reddish-brown oil which was dissolved in CDCl3 (0.5 mL) for NMR sampling. Each NMR spectrum was analysed according to the method reported below (Section 2.7) and the results were graphed using Origin 2018 software.
A second Schlenk tube (Flask 2) was charged with 2,4-dibromopyridine 1 before being evacuated and backfilled with N2 three times. THF (1.3 mL) was added and the resulting clear solution was stirred at room temperature.
After the specified premixing time, the contents of Flask 1 were transferred via cannula to Flask 2. The resulting mixture was allowed to equilibrate for 1 minute with stirring after which time phenylmagnesium bromide solution (5, 1.2 mL, 1.0 M in THF => total volume of 5 mL) was added in a rapid injection. The subsequent solution was stirred for one hour.
A sample (125 µL) was taken via a syringe and quickly quenched by addition to a vial containing 100 µL of NH4Cl solution (aq., sat.). EtOAc (2 mL) was added to the sample and the resulting mixture was vigorously shaken. The EtOAc was removed using a pipette and the aqueous layer was extracted with another aliquot of EtOAc (2 mL).
The combined organics were filtered through a MgSO4 plug (ca. 1 cm depth) which was washed through with EtOAc (1 mL). The filtrate was concentrated in vacuo to afford a yellow oil which was dissolved in CDCl3 and subjected to 1 H NMR analysis (see Section 2.7).

General Workup Procedure for SMCC Reactions
The above cross-coupling reactions (from Sections 2.1-2.5) were worked-up using the following procedure. After the reaction time, the reaction solution was quenched using a of NH4Cl (sat. aq). The organics were then extracted using EtOAc (4 × 10 mL) and combined before drying over MgSO4, filtered and subsequently concentrated in vacuo. The resulting residue, which generally appeared as a reddish-brown oil, could be purified by column chromatography (SiO2) using a hexane or petroleum ether/EtOAc solvent system. See Section 3 for purification details for specific compounds.

1 H NMR Analysis of NMR Samples of Cross-coupling Reactions
The reaction conversion and site-selectivity of cross-coupling reactions at 2,4-dibromopyridine 1 were determined using a 1 H NMR spectroscopy-based assay of crude reaction samples (unless otherwise stated). Crude reaction samples were prepared according to the relevant method. Products were identified from the crude reaction mixtures by reference to literature published data or, if the product was not reported, by comparison to characterisation data obtained for the isolated product. The pyridyl-H6 protons of the starting materials and products were generally well-resolved, due to the significant deshielding (proximity to electronegative N), thus, integration of these protons was generally used to determine the relative amounts of starting material and products; C2Ar, C4Ar and diaryl (see Figure S1 for a typical example of how this integration was done for each reaction).  The above method was validated using an internal standard, 1,3,5-trimethoxybenzene. 1,3,5-Trimethoxybenzene The products could then be quantified against the internal standard as follows (Table S8), providing confidence in the assay as a validated quantitative method.
The crystalline product could therefore be isolated for further characterisation (vide infra) in a 26% yield of isolated OAC2-Br product (unoptimized crystallisation). Upon subjecting one such crystal to XRD analysis, the solidstate structure was confirmed as being that of the OAC2-Br oxidative addition product ( Figure S4). aromatic resonances, at δP 6.65 and 6.23 ppm (Right, Figure S3), which were assigned as -H3 and -H5 protons on the 4-brompyridyl ligand based on their coupling constants and relative integrations. This upfield-shifting of these proton resonances is likely a result of the interaction of the aromatic ring currents associated with the proximal phenyl moieties. Such an interaction is evident in the crystal structure ( Figure S4), which shows an interfacial πstacking interaction -the 4-bromopyridyl ligand is sandwiched between two phenyl groups from the transconfigured triphenylphosphine ligands, providing further support that the crystal obtained (XRD) is the major species observed by 1 H NMR spectroscopic analysis (in solution). See Section 4 for X-Ray Crystallographic Data (ijsf1805).

H NMR analysis of the degraded solution of Pd3(OAc)6 and 3 PPh3
Acetic anhydride was detected by comparison of the 1 H NMR spectrum of the post-reaction mixture with the 1 H NMR spectrum containing an authentic sample, both in THF-d8 ( Figure S6). Acetic acid was also detected as a biproduct of the process and they appear in the post-reaction solution in an apparent molecular ratio (Ac2O:AcOH)  4 was not isolated in preparative form from solution due to its instability (vide supra and discussion in the main paper). The following data is presented for its characterisation in solution after direct reaction between Pd3(OAc)6 and 3 equivalents of PPh3, in THF-d8 as highlighted above.

Isolation of single crystal of 4
A Schlenk tube was charged with Pd3(OAc)6 (6.7 mg, 0.01 mmol) and 3 equivalents of PPh3 ({7.9 mg, 0.03 mmol}: Pd:P = 1:1). The Schlenk tube atmosphere was evacuated and backfilled with N2 before being put on ice. THF (2.5 mL; pre-cooled to 0 o C) was added and the resulting mixture was stirred at 0 o C for 5 minutes, appearing as a brownish-red solution. A sample (0.5 mL) was taken via a syringe and added to a J. Youngs-type NMR tube under Schlenk conditions, on ice. The reaction mixture was then layered with pre-cooled hexane and swiftly stored at - This crystal data is presented in the main paper and in Section 4 of this document. S18

SMCC of 1 with 4-Fluorophenylboronic acid (2a)
Products could be identified from crude reaction mixtures ( Figure S7). Characterisation data matched that reported in the literature, 3 and/or with data collated for (novel) isolated compounds reported vide infra.

4-Bromo-2-(4-fluorophenyl)pyridine (3aC2-Ar) 3
Compound characterisation data agrees with that previously reported in literature. 3 It was isolated from the reaction mixture using flash chromatography (SiO2) with a hexane/EtOAc (95:5) solvent system, which was run S20 with a gradient, starting from neat hexane. Yield from the model SMCC (Section 2.5), catalysed by Pd3Cl2 = 21 mg (8%). Appeared as a colourless oily film.  13  Compound characterisation data agrees with that previously reported in literature. 6 It was isolated from the reaction mixture using flash chromatography (SiO2) a hexane/EtOAc (95:5) solvent system, which was run with a gradient, starting from neat hexane. Yield from an SMCC reaction (Section 2.5), catalysed by Pd3Cl2 = 50.0 mg (19 %). Appeared as a colorless powder.  Products could be identified from crude reaction mixtures ( Figure S8). Characterisation data matched that reported in the literature, 3 and/or with data collated for isolated compounds reported vide infra.

4-Bromo-2-(4-anisyl)pyridine (3bC2-Ar) 3
Compound characterisation data agrees with that previously reported in literature. 3   Products could be identified from crude reaction mixtures ( Figure S9). Characterisation data matched that reported in the literature, 3 and/or with data collated for isolated compounds reported vide infra.

2-Bromo-4-phenylpyridine (3cC4-Ar) 3
Compound characterisation data agrees with that previously reported in literature. 3 It was isolated from the

4-Bromo-2-phenylpyridine (3cC2-Ar)
Reaction data for this compound, characterised as part of the crude post-reaction mixture, matched that previously reported in literature (see Figure S9 above).

2,4-Bis-phenylpyridine (3cdiaryl)
Reaction data for this compound, obtained as part of the crude post-reaction mixture, matched that previously reported in literature (see Figure S9 above

SMCC of 1 with 4-Tolylboronic acid (2d)
Products could be identified from crude reaction mixtures ( Figure S10). Characterisation data matched that reported in the literature, 3, 7-8 and/or with data collated for isolated compounds reported vide infra.

4-Bromo-2-(4-tolyl)pyridine (3dC2-Ar)
Compound characterisation data agrees with that previously reported in literature. 3  Compound characterisation data agrees with that previously reported in literature. It was isolated from the reaction (Section 2.1, catalyzed by Pd3Cl2) mixture using flash chromatography with a hexane/EtOAc (95:5 v/v) solvent system, which was run with a gradient starting from neat hexane. Yield = 24.6 mg (9%

SMCC of 1 with 4-Chlorophenylboronic acid (2e)
Products could be identified from crude reaction mixtures ( Figure S11). Characterisation data matched that reported in the literature, 7, 10-11 and/or with data collated for (novel) isolated compounds reported vide infra.  10 Compound characterisation data agrees with that previously reported in literature. It was isolated from the reaction mixture (Section 2.5, catalyzed by Pd3Cl2) using flash chromatography (SiO2) with a hexane/EtOAc (98:2 v/v) solvent system, which was run with a gradient starting from neat hexane. Yield from reaction catalysed by

4-Bromo-2-(4-chlorophenyl)pyridine (3eC2-Ar) 11
Compound characterisation data agrees with that previously reported in literature. 11 It was isolated from the reaction mixture (Section 2.5) using flash chromatography (SiO2) with a hexane/EtOAc (98:2 v/v) solvent system, which was run with a gradient. Yield from reaction catalysed by Pd3Cl2 = 34.7 mg (12.9 %). Appeared as a colorless powder. 1   Compound characterisation data agrees with that previously reported in literature. 7 It was isolated from the reaction mixture (Section 2.5) using flash chromatography with a hexane/EtOAc (98:2) solvent system, which was run with a gradient starting from neat hexane. Yield from reaction catalysed by Pd3Cl2 = 22.9 mg (8%). Appeared as a colorless powder.  Products could be identified from crude reaction mixtures ( Figure S12). Characterisation data matched that reported in the literature, 7,12 and/or with data collated for isolated compounds reported vide infra.

XRD analysis (ijsf1803; CCDC 2060853):
Crystal was grown by vapour diffusion of pentane on a dichloromethane solution of 3fC4-Ar.

Figure S13 Structure obtained from X-ray diffraction analysis of a single crystal of 2-bromo, 4-(4trifluoromethylphenyl)pyridine (3fC4-Ar).
The exact crystal run (by XRD) was subjected to 1 H NMR spectroscopic analysis (700 MHz) which matched data of that of the isolated material (vide supra). This confirms conclusively the identity of the NMR characterised species as the C4-arylated product. Thus, X-ray crystallographic analysis shows distortion of the pyridine ring (Table S9 &   Table S10).

4-Bromo-2-(4-trifluoromethylphenylphenyl)pyridine (3fC2-Ar)
To our knowledge, data for this compound was not previously reported within literature. It was isolated from the crude reaction mixture, catalysed by Pd3Cl2 (Section 2.5) using PTLC (SiO2) with a hexane/EtOAc (98:2) solvent system. Yield from reaction catalysed by Pd3Cl2 = 3.7 mg (1 %). Multiple solvent runs were needed to allow separation from the starting material. Appeared as a colorless powder.

Assessment of the effects of tris-imid.3Br 6 in a SMCC reaction between 1 and 2b
Tris-imidazolium tribromide 6 was applied to the benchmark SMCC conditions (see conditions and results below), catalyzed by our optimized catalyst precursor system, Pd(OAc)2/1 PPh3. We noted a marked rise in site-selectivity at 1, exhibiting a 3bC4-Ar:3bC2-Ar ratio of 17.6:1, with a relatively low formation of 3bdiaryl product. before being combined and dried over MgSO4, filtered and subsequently concentrated in vacuo. The resulting residue, which appeared as a reddish-brown oil was adsorbed to silica and purified by flash chromatography (SiO2) using a Hexane or petroleum ether/EtOAc (85:15) solvent system to give the product which appeared as a colourless oil (550 mg, 69%).
Step 2 Ullman Coupling at 3b4-Ar: Adapted from a procedure published by Zhao et al. 17

X-Ray Crystallography
Diffraction data were collected at 110 K on an Oxford Diffraction SuperNova diffractometer with Cu-K radiation ( = 1.54184 Å using an EOS CCD camera. The crystal was cooled with an Oxford Instruments Cryojet. Diffractometer control, data collection, initial unit cell determination, frame integration and unit-cell refinement was carried out with "Crysalis". 18 Face-indexed absorption corrections were applied using spherical harmonics, implemented in SCALE3 ABSPACK scaling algorithm. 19 OLEX2 20 was used for overall structure solution and refinement . Within OLEX2, the algorithm used for structure solution was "ShelXT dual-space". 21

Refinement Special Details
See Figure S13 for single crystal structure image. The fluorine atoms within the CF3 group were disordered and modelled in two positions with refined occupancies of 0.820:0.180 (18), the ADPs of the fluorine atoms were restrained to be approximately isotropic.

Refinement Special Details
The asymmetric unit contained half a molecule, the other half being generated by a mirror plane.

Refinement Special Details
The compound was prepared and crystallised as detailed in Section 3.2. See Figure S14 for single crystal structure image. The structure was disordered about a mirror plane parallel to the ac plane at b=0.25. In the central ring, C3 and C4 occupied a common site for both conformations.