New Sustainable Pd(II)/Fe(III) Catalytic System Very Efficient in the Hydromethoxycarbonylation of 1-octene

In order to increase the catalytic activity of the Pd(II)-complexes in the alcoxycarbonylation of 1-olefins, the presence in solution of a strong Brönsted acid together with a phosphine ligand is usually required. From an industrial point of view, such two additives, however, influenced the sustainability of the process mainly caused by the related corrosion of the reaction vessel and the toxicity of the phosphine ligand. In this paper, the methoxycarbonylation of 1-octene has been efficiently carried out by using the [Pd(TsO) 2 (PPh 3 ) 2 ]/FeCl 3 catalyst system. The catalyst is very active also without addition of PPh 3 , reaching the TON of ca . 1300 (mol/mol) in 2 hours by using the Fe(III)/Pd(II) = 400/1 (mol/mol) system. Such value is higher than the values obtained by using the traditional strong Brönsted acid TsOH or other Lewis acid such as AlCl 3 , in the presence of PPh 3 . The reaction takes place with the formation of two isomeric products of linear and branched structures with a yield towards the linear isomer of ca. 80%.

Although several transition metal complexes can be used as catalyst (Ru, Pd, Ni or Co complexes [23,[31][32][33][34][35][36]), the most promising appear the Pd(II) complexes, in terms of conversion, selectivity and process conditions, as to allow industrial applications.As an example of hydroalkoxycarbonylation, which has found industrial application, we can mention the Alpha process (developed by Lucite International) [37,38], which produce methyl methacrylate through the reaction of formaldehyde with methylpropionate.The latter is produced, in the first step of such process, via hydromethoxycarbonylation of ethylene by using a Pd(II)-phosphine complex as catalyst precursor.
The choice of the phosphine ligand plays a fundamental role in determining the catalytic activity of the Pd(II)complex, but the co-presence in solution of an excess of the same ligand and of a Brönsted acid, is almost always needed, in order to obtain the performance required.In fact, the appropriate acidic strength and the coordinative properties of the Brönsted acid (e.g., p-toluenesulfonic acid, TsOH) favor the in situ formation of active Pd-hydride species [26-29, 39, 40], whose decomposition to inactive Pd-metal is efficiently avoided by the presence in solution of the excess of the phosphine ligand (PPh 3 ) [9,[39][40][41][42][43] (see Scheme 1).However, from an industrial point of view, we have also to consider the possible drawbacks and environmental impacts relate to such reaction.For instance, the use of hazardous chemicals and toxic solvents, the lack of substrate selectivity, excessive consumption of energy and possible corrosion of the equipment, which can influence the sustainability of the process [3,4,8,9,44].Among these points, some attention should be paid to the addition of strong Brönsted acids, which could lead to corrosion problems in the reaction vessel and in the equipment, and to the addition of extra phosphine ligands, which could increase the toxic-related problems, e.g. in the food industry.To address such Brönsted acid related problems, the Lewis acids are increasingly used, as they have been shown to be effective for a variety of hydroalkoxycarbonylation protocols [45][46][47][48][49].In this regard, in a previous paper, some of the authors have studied the effectiveness of AlCl 3 as substitute of the p-toluenesulfonic acid (TsOH) in the hydroethoxycarbonylation of 1-olefins [48][49][50][51].
In this paper the use of Lewis acids has been further investigated, funding that the [Pd(TsO) 2 (PPh 3 ) 2 ]/FeCl 3 (TsO = p-tosylate) catalyst system led to the best results with 1-octene in methanol without adding toxic PPh 3 .The use of cheap and very available FeCl 3 , together with the absence of toxic PPh 3 in solution, represent an interesting result in terms of sustainability of the reaction.The comparison with TsOH and AlCl 3 has been also proposed and discussed.

Equipments and characterization
The catalyst precursors have been weighted on a Sartorious Micro balance (precision 0.001 mg).
GC/MS analyses were performed on a MS Agilent apparatus 5975C Model, interfaced with an Agilent chromatograph 7890 A Model equipped with a HP5 column (30 m × 0.25 mm × 0.25 μm, oven: 318 K (3 min) to 523 K at 15 K/min).
Fourier transform infrared (FTIR) spectra were recorded on a Nicolet Magna 750 instrument in KBr powder.All the NMR spectra were recorded on a Bruker Avance 300 spectrometer.

Catalytic reactions
The catalysis has been carried out in a batch reactor of ca.60 mL provided with a magnetic stirrer.In a typical experiment, 1.50 10 −3 mmol of Pd(II) complex (for instance 1.46 mg of [Pd(TsO) 2 (PPh 3 ) 2 ]), together with the excess of triphenylphosphine (PPh 3 ) and the acid in a molar ratio Pd(II)/PPh 3 /AlCl 3 = 1/12/100 (mol/mol/mol)) were added to 20 mL of solvent (methanol).In order to avoid catalyst deactivation, due to the presence of air, the reactor was carefully flushed with CO at room temperature with stirring.The reactor was then heated up to 393 K in ca. 10 min without stirring under a 0.5 MPa of CO pressure.At 393 K the pressure has been then adjusted to the desired value (typically 5.0 MPa of CO) and, while stirring, maintained constant by continuously supplying the CO from a reservoir throughout the experiment (for instance 3 h).At the end of the experiment, the reactor was quickly cooled and carefully depressurized at room temperature.
The reaction products have been detected and quantified by the GC and GC-MS analysis.
All the experiments have been repeated at least twice and the error analysis confirmed the reproducibility of the data with an error of ±1%.
Therefore, focusing on 1-octene, we have tested the influence of different Lewis acids (Table 1 entries 7-14) in comparison to TsOH (entry 6).Among these, the FeCl 3 leads to an interesting result both in terms of conversion and yield to linear ester (74% of conversion; 75% of linear esters; entry 9), also considering that it is cheap and very available.
Moreover, the yield to linear isomer obtained with the [Pd(TsO) 2 (PPh 3 ) 2 ]/FeCl 3 /PPh 3 catalyst system (Table 1, entry 4) is similar to the values reported in the literature [50,51].According to literature, Table 2 shows that the weakly coordinating ligands lead to the best catalytic activity (Table 2 entries 1-4).In addition the influence of different phosphine ligands, in terms of relative nucleophilicity and steric properties, have been also tested substituting the phosphine in both the Pd(II) complex and in the solution as excess (Table 2 entries 5-10).
The results confirm that the optimum combination of the stereo-electronic effects are obtained with PPh 3 , according to the conclusion that more basic ligands disfavors the coordination of the nucleophilic alkene to the metal centre, while with the less basic phosphines the alkene binds too strongly to the metal disfavoring the catalysis.
On the basis of these results the catalytic activity of the [Pd(TsO) 2 (PPh 3 ) 2 ]/FeCl 3 /PPh 3 system has been further investigated.

On the catalytic activity of the [Pd(TsO) 2 (PPh 3 ) 2 ]/ FeCl 3 /PPh 3 system
The Fig. 1 shows the influence on the catalytic activity of FeCl 3 concentrations, compared to TsOH and AlCl 3 .In all the cases, the addition of a co-catalyst is needed and the catalytic activity linearly increases by increasing the co-catalyst/Pd molar ratio following similar trends, even if some differences can be highlighted.For instance, at low concentration (co-catalyst/Pd = 10/1 mol/mol), the more efficient was TsOH (TON = 650 moles of esters/moles of Pd) respect to AlCl 3 (TON = 542) and FeCl 3 (TON = 466).On the other hand, at high concentration (co-catalyst/Pd > 400/1 mol/mol), the FeCl 3 led to the best performance (TON = 1200 moles of esters/moles of Pd) compared to TsOH (TON = 1150) and AlCl 3 (TON = 1100).

Influence of the addition of PPh 3
The PPh 3 /Pd molar ratio was also found to have a significant role in catalytic activity.The main role of PPh 3 , added to the [Pd(TsO) 2 (PPh 3 ) 2 ] catalyst, is to avoid the formation and the precipitation of inactive Pd metal (see introduction [26,27,41], increasing therefore the catalyst performance.According to this, the Fig. 2 shows that, in the presence of TsOH, the TON increases with the PPh 3 concentration, reaching a plateau value of ca.800 (mol/mol) at PPh 3 /Pd >8/1.However, by using AlCl 3 instead of TsOH, the TON increases up to a maximum value of ca.700 (mol/mol) at PPh 3 /Pd higher than 12/1 (mol/mol), after then it quickly decreases for PPh 3 /Pd > 100/1 (mol/mol).On the contrary, by using FeCl 3 , the TON linearly decreases by increasing the PPh 3 /Pd molar ratio, throughout the range studied.This suggests that the increase of the PPh 3 concentration favors the formation of phosphine complexes also with Fe [56][57][58][59][60], which probably are less or not active as co-catalysts.According to this, the Fig. 2 shows that FeCl 3 was more active in absence of PPh 3 , leading to a TON of 630 (mol/mol).Such TON can by readily increases by increasing the FeCl 3 /Pd molar ratio, reaching for instance ca.1300 (mol/mol) by using the FeCl 3 /Pd = 400/1 (mol/mol).
In all the experiments in Figs. 1 and 2, no by-products were detected (selectivity to esters >99%) having a yield to linear ester of 80-82%, by using FeCl 3 and TsOH, and 65-70% by using AlCl 3 .

On the reaction mechanism
Although the mechanism of hydroalcoxycarbonylation of olefins in the presence of Brönsted acid is widely studied in the literature, we can do some supposition on the role played by FeCl 3 as substituted of the Brönsted acid.Under the reaction conditions, the tosylate ligands in the Pd(II) precursor (weakly coordinating) are readily substituted by the solvent forming a cationic complex in which they are the counter-anions [42] (a in Scheme 2).The active species for the hydromethoxycarbonylation (b and c in Scheme 2), can easily form through reactions 2 and 3. Species b and c start two possible cycles to form the esters (isomers) together with the active species, which restart a new cycle.However, it is widely reported that the redox deactivation of the Pd(II) active species can readily occur forming less/ not active Pd(0) species or Pd metal [42], with a consequent progressive decrease of the catalytic activity in the absence of an excess of Brönsted acid.Therefore, in light of the results reported In the present manuscript, it is plausible to suppose that the addition of FeCl 3 forms a redox couple with Pd(0), which gives Pd(II) active species again [61] (Scheme 2, reaction 5).Therefore the catalysis occurs without formal reduction to Pd(0) during the catalytic cycle (fast

Conclusion
The FeCl 3 results an effective and interesting substitute of the p-toluenesulfonic acid and AlCl 3 as promoter for the hydromethoxycarbonylation of 1-octene.The reaction takes place with the formation of two isomeric products of linear and branched structures without by-products.Under the best reaction conditions tested, the TON (sum of isomeric esters of linear and branched structures) reaches a value of ca.1300 (mol/mol) by using the FeCl 3 = 400/1 molar ratio, with a yield towards the linear isomer of ca.80 %.Furthermore, the use of FeCl 3 as co-catalyst allows to carry out the phosphine-free reaction, avoiding the toxicity related problems.

Table 2
Influence of different Pd(II) precursor in the