Competing Mechanisms in Palladium-Catalyzed Alkoxycarbonylation of Styrene

Palladium-catalyzed carbonylation is a versatile method for the synthesis of various aldehydes, esters, lactones, or lactams. Alkoxycarbonylation of alkenes with carbon monoxide and alcohol produces either saturated or unsaturated esters as a result of two distinct catalytic cycles. The existing literature presents an inconsistent account of the procedures favoring oxidative carbonylation products. In this study, we have monitored the intermediates featured in both catalytic cycles of the methoxycarbonylation of styrene PhCH=CH2 as a model substrate, including all short-lived intermediates, using mass spectrometry. Comparing the reaction kinetics of the intermediates in both cycles in the same reaction mixture shows that the reaction proceeding via alkoxy intermediate [PdII]-OR, which gives rise to the unsaturated product PhCH=CHCO2Me, is faster. However, with an advancing reaction time, the gradually changing reaction conditions begin to favor the catalytic cycle dominated by palladium hydride [PdII]-H and alkyl intermediates, affording the saturated products PhCH2CH2CO2Me and PhCH(CO2Me)CH3 preferentially. The role of the oxidant proved to be crucial: using p-benzoquinone results in a gradual decrease of the pH during the reaction, swaying the system from oxidative conditions toward the palladium hydride cycle. By contrast, copper(II) acetate as an oxidant guards the pH within the 5–7 range and facilitates the formation of the alkoxy palladium complex [PdII]-OR, which favors the oxidative reaction producing PhCH=CHCO2Me with high selectivity. Hence, it is the oxidant, rather than the catalyst, that controls the reaction outcome by a mechanistic switch. Unraveling these principles broadens the scope for developing alkoxycarbonylation reactions and their application in organic synthesis.


Experimental details
Mass spectrometric experiments were performed on Thermo Scientific LTQ XL, Finnigan LCQ Deca XP mass spectrometer, or Bruker TIMSTOF, all equipped with electrospray ionization (ESI) sources. 1,2General conditions for LTQ and LCQ were as follows: sheath gas 5-40 arbitrary unit, auxiliary gas 0-10 arbitrary unit, capillary temperature 180-220 ℃, spray voltage 3-5 kV, capillary voltage -20 to 30 V, and tube lens -40 to 50 V.The energy-resolved collision-induced dissociation (CID) experiments were performed on LCQ Deca mass spectrometer with an ESI source.The negative mode collision energies in the LCQ ion trap were calibrated based on the measurements of dissociation energies of carboxylate anions (trifluoroacetate, dichloroacetate, trichloroacetate and benzoate). 3,4The mass selected ions were collided with the He buffer gas with an activation time of 30 ms and activation q = 0.25.The complexes were repeated 2-4 times to determine the standard deviation of determined appearance energies (AEs).

General procedure for the offline sampling of reaction mixture:
To a solution of PdCl2 (100 µM) in a mixture of acetonitrile (1 ml) and methanol (1 ml), p-benzoquinone (1 mM) was added.The clear solution was pre-stirred under the carbon monoxide (CO) atmosphere (using a balloon) for 2-5 minutes, and styrene (1-5 mM) was added.The reaction mixture was then stirred at 40 ℃ for 20-80 minutes.Aliquots of the reaction mixture were then filtered (using syringe filter) and analyzed on ESI-MS.The substrate and/or solvent were substituted for labeling experiments as indicated.
General procedure for the online sampling of reaction mixture i.e. pressurized sample infusion-electrospray ionization-mass spectrometry (PSI-ESI-MS) 5,6 monitoring: To a solution of PdCl2(100 µM) in a mixture of acetonitrile (1 ml) and methanol (1 ml), p-benzoquinone (1 mM) was added.The clear solution was pre-stirred under the carbon monoxide (CO) atmosphere (using a balloon) for 2-5 minutes.This stirred reaction mixture was directly sprayed into the ESI-MS inlet via silica capillary and N2 or CO overpressure.When required using a heating metal block, the desired temperature was achieved.Substrate styrene (1-5 mM) was injected during the acquiring MS data to monitor the evolution of the reaction intermediates.The substrate and/or solvent were modified for the labeling experiments as indicated.

GC-TOF measurements:
We performed 1:10 dilution of the reaction mixture at specified intervals with dichrolomethane (DCM), followed by extraction with water, resulting organic layer was dried over sodium sulphate, filtered by syringe filter and injected to the GC-TOF.The Agilent 7890A GC-TOF used was fitted with electron ionization (EI) source and HP-5MS column (30m x 0.25mm x 0.25µm).Method employed (15 min run time) started with the oven temperature at 100 ℃ (1 min hold), 20 ℃ increment until 320 ℃ (3 min hold).Detector was set at 2050 V and split ratio of 10.0:1 was applied.
pH measurements: As the reaction solvent(/s) i.e. acetonitrile and methanol are miscible with water, we prepared 1:1 dilution of the reaction mixture at selected interval with Milli-Q water, followed by mixing and syringe filtration to remove black precipitate.The resulting solution was tested with universal pH paper for qualitative pH estimation.Below are the zoomed experimental spectrum sections (top) plotted against the calculated spectrum (bottom) for isotopic pattern analysis of the palladium complexes.Below are the zoomed experimental spectrum sections (top) plotted against the calculated spectrum (bottom) for isotopic pattern analysis of the palladium complexes.Below are the zoomed experimental spectrum sections (top) plotted against the calculated spectrum (bottom) for isotopic pattern analysis of the palladium complexes.Calibration of the ion trap of the LCQ-Deca electrospray ionization mass spectrometer in the negative mode using specified carboxylate ions of known bond dissociation energies.The appearance energies (AEs) were obtained by sigmoidal fitting of the fragment intensity vs applied collision energy.The DFT calculations were performed using the B3LYP 7 functional with the D3 dispersion correction 8 as implemented in Gaussian 16. 9 All the calculation were carried out with the 6-311+G** basis set 10,11 and SDD 12,13 on palladium.The geometries were fully optimized by verifying that there were no imaginary frequencies.

Collision cross section (CCS) calculation
The DFT calculations were performed using the B3LYP 7 functional with the D3 dispersion correction 8 as implemented in Gaussian 16. 9 All the calculation were done with the 6-311+G** basis set 10,11 and SDD 12,13 on palladium.The geometries were fully optimized by verifying that there were no imaginary frequencies.The CCS calculation were performed using the coordinates of optimized geometries in Collidoscope 14 'trajectory method' based CCS calculator using N2 as the collision gas.Ion mobility separation experiments were performed on Bruker timsTOF, the mass and mobilities were calibrated using standard calibration solution. 15Using the DataAnalysis and Bruker Compass mobility calculator, the experimental CCS was determined.We note that the calculated CCS for the structure (A) and (B) is very close with only ~ 2.6 Å 2 difference, If both the isomers are present then they can be separated only based on the resolution of ion mobility separation technique used.In our case, the major isomer with ~ 140 Å 2 was detected which we assign to structure (A).It must be noted that structure (A) is 17 kJ mol -1 lower in energy when compared to structure (B).Based on the comparison of experimental and calculated CCS along with the relative energies the major isomer detected is assigned to structure (C).Experiments with no oxidant in the ultra resolution also showed a minor isomer with comparatively small CCS, which could be structure (B).Here we note that the lowest energy structure has huge difference in CCS with the experimentally obtained CCS.Based on the comparison of experimental and calculated CCS along with the relative energies, the major isomer detected is assigned structure (C).The DRL fitting to obtain t1/2 in red did not fit the experimental data properly.Alternatively, using double exponential non linear curve fitting i.e.ExpGrow2 (dark red), gave 2 processes.As in this experiment the alcohol is added, this alcohol will first form Pd alkoxycarbonyl precursor and then the [PdCl2(styrene,COOCH3)] -intermediate.

Figure S34:
Snapshots of online monitoring of the reaction mixture of PdCl2(100 µM), p-benzoquinone (500 µM) and styrene (500 µM) in acetonitrile and methanol (1:1) stirred under CO atmosphere at room temperature.Initially the alkoxy intermediates are observed, followed by hydride intermediates.At the end the intermediates of hydride cycle prevails.EI-MS spectrum of peak observed at retention time '6.33' and the proposed scheme for its formation from the dicarbonylated product.
• We observe that the reaction with p-BQ shows unsaturated oxidative carbonylation product at '5.26' at 30-60 min.When compared to other products, this unsaturated product does not grow after long time.This is in-line with our findings that the alkoxy cycle could be observed only at the start of the reaction.• Over long time when the pH becomes slightly acidic, then the hydride cycle becomes dominant leading to the saturated products and is reflected in the chromatogram.
• Similarly, using stoichiometric PdCl2 with no additional oxidant showed acidic pH and resulted in the detection of the saturated carbonylated products of the hydride cycle after 2 hr at 40 ℃.

Effect of the reaction conditions on the observation of the intermediates
• 1,1′-Bis(diphenylphosphino)ferrocene (dppf) as the bidendate phosphine ligand       ESI-MS spectrum of PdCl2, copper acetate monohydrate, tetrabutylammonium bromide in acetonitrile under CO atmosphere, showing various Pd and copper complex/clusters.This is pre-stirred mixture according to literature before styrene addition. 16Cl2_BQ_ACN_MeOH_COatm_clstyrene #5-116 RT: 0.   It is important to note that -OMe containing complexes were not detected in this case when compared to oxidant copper acectate.• Palladium and copper speciation under the reaction conditions Stock solutions of palladium chloride 2.6 mM in CH3CN, Cu(OAc)2.H2O 18.7 mM in CH3CN, styrene 144 mM in CH3CN were prepared.The palladium chloride and copper acetate monohydrate were sonicated for 20-30 min.To 1:1 solvent mixture of CH3CN:MeOH, 200 µM of PdCl2 and 650 µM Cu(OAc)2.H2O was added.This mixture was stirred at room temperature under the balloon of carbon monoxide (CO) and oxygen.The reaction mixture was online monitored and at 1 st min excess of styrene (10 mM) was added.Resulting reaction mixture was monitored for an hour on ESI-TOF.• At the start, we observed the Pd(II) complexes as depicted in golden yellow to orange shades.

• Copper acetate monohydrate as oxidant and in the presence of tertabutyl ammonium bromide
• After styrene addition (dotted grey line indicating the addition), the red curve shows that the Pd complex containing styrene increases sharply and, over time, decreases.• The purple shades curves indicated di copper complexes with bridging ligands.It must be noted that the dicopper complexes with methoxyl as the bridging ligand are only detected after the species containing the styrene has disappeared.During this time, the Pd(II) (grey curve) is generated simultaneously.• Bimetallic Pd Cu complexes show the interplay of the bridging ligands.Initially, complexes with chloro bridging ligands are observed (cyan), followed by mixed choro with acetato (green), and finally, towards the end, only the complexes with the acetato bridging ligands are observed (in blue).

Figure S49:
Plot snapshots of ESI-MS online monitoring spectra of the reaction mixture of under the O2 atmosphere at room temperature a) initial spectrum of PdCl2 in CH3CN, b) on addition of copper acetate monohydrate, c) on the addition of methanol and d) spectrum after the addition of styrene in 20 minutes.
• At the start, we observed the Pd(II) complexes in both positive and negative modes.
• After the copper addition, followed by methanol, we initially observed copper(II) complexes in the negative mode.• On the addition of styrene and as the reaction proceeds, copper(I) complexes are detected in both modes.• We observed the [(PdCl2)nCl]⁻ complexes and their adducts with Pd 0 .Because of the low solubility of PdCl2 in methanol, we had to sonicate the solution for ~20 min.During this time, some palladium was reduced.Palladium(II) complex can react with the methanol ligand by -hydrogen elimination, leading to palladium hydride and ultimately to palladium reduction.

Figure S18 :
Figure S18: Energy resolved CID and extrapolation of fragmentation onset to determine the appearance energy (AE) of CO in the intermediate [PdCl3(CO)] -with m/z 239.

Figure S19 :
Figure S19: Energy resolved CID and extrapolation of fragmentation onset to determine the appearance energy (AE) of acetolactone in the intermediate [PdCl2(COOCH3)] -with m/z 235.

Figure S20 :
Figure S20: Energy resolved CID and extrapolation of fragmentation onset to determine the appearance energy (AE) of CO in the intermediate [PdCl2(COOCH3,CO)] -with m/z 263.

Figure S22 :
Figure S22: Energy resolved CID and extrapolation of fragmentation onset to determine the appearance energy (AE) of HCl in the intermediate [PdCl2(H,styrene)] -with m/z 281.

Figure S23 :
Figure S23: Energy resolved CID and extrapolation of fragmentation onset to determine the appearance energy (AE) of CO in the intermediate [PdCl2(H,styrene,CO)] -with m/z 309.

Figure S26 :
Figure S26: Energy resolved CID and extrapolation of fragmentation onset to determine the appearance energy (AE) of methyl cinnamate in the intermediate [PdCl((styrene-H)COOCH3)] - with m/z 303.
Figure S27: a) ESI-TOF spectra of the reaction mixture of PdCl2 (200 µM) in a mixture of acetonitrile (1 ml) and methanol (0.5 ml), p-benzoquinone (1 mM), and styrene (5 mM) under the CO atmosphere at 40 ℃ for 30 min, black precipitation filtered by syringe filter; b) c) d) and e) Mobilogram of specified m/z with inset showing ultra resolution mobilogram from 0.65 to 0.85 range.Dashed line indicate isomer appearing due to fragmentation.
Figure S28: a) ESI-TOF spectra of the reaction mixture of excess of PdCl2 and styrene in CH3CN:MeOH under the CO atmosphere at 40 ℃ for 30 min, black precipitation filtered by syringe filter; b) c) d) and e) Mobilogram of specified m/z with inset showing ultra resolution mobilogram from 0.65 to 0.85 range.Dashed line indicates isomer appearing due to fragmentation.

Figure S30 :
Figure S30:IRMPD of the mass selected [PdCl2(H,styrene)] -(top experimental spectrum; black line represents m/z 281 for styrene while the red line represents corresponding complex with D3-styrene) with the theoretical spectra of possible isomers (styrene-grey and D3-styrene-light red).

Figure S31 :
Figure S31:IRMPD of the mass selected [PdCl2(H,styrene,CO)] -(top experimental spectrum; black line represents m/z 309 for styrene while the red line represents corresponding complex with D3-styrene) with the theoretical spectra of possible isomers (styrene-grey and D3-styrene-light red).

Figure S35 :
Figure S35: GC-MS chromatograms of the sample from the reaction mixture (a) PdCl2 (10 mM), pbenzoquinone (40 mM), styrene (40 mM) in acetonitrile and CH3OH (1:1) ratio under CO atmosphere at 40 ℃ after i) 30 minutes, ii) 60 min, iii) 2 hr and iv) 4 hr; (b) PdCl2 (10 mM), Cu(OAc)2.H2O (40 mM), styrene (40 mM) in acetonitrile and CH3OH (1:1) ratio under CO and O2 atmosphere at 40 ℃ after i) 30 minutes and ii) 60 min; (c) Structure are assigned based on comparison of EI spectrum with the NIST library.[citation]Except for the '6.33' which is not available in the NIST library and is proposed based on the m/z and fragmentation observed as a subsequent intramolecular lactonization from the dicarbonylated product.
Figure S36:EI-MS spectrum of peak observed at retention time '6.33' and the proposed scheme for its formation from the dicarbonylated product.

Figure S47 :
Figure S47:Snapshots of ESI-TOF online monitoring spectra of the reaction mixture of PdCl2 and copper acetate monohydrate in CH3CN:MeOH under the CO and O2 atmosphere at room temperature a) initial spectrum, b) on addition of styrene and c) spectrum after 30 minutes, with zoomed section.