The production of aromatic esters represents the core business of several chemical industries, being intermediates in the manufacture of important products such as agrochemicals, pharmaceuticals and others [1–3]. Although the classical synthesis of esters generally requires the reaction of a carboxylic acid precursor in the presence of the suitable nucleophile [4–8], several alternative methods are effective, including the alkoxycarbonylation of aryl halides [4, 9–11].
Despite the well-known toxicity of carbon monoxide (CO), it is widely used as an inexpensive and readily available C1 source and in an overall sense it can be considered an environmentally friendly source . In the presence of a suitable catalyst, CO can be incorporated into a variety of organic compounds in its entirety without producing any undesirable byproduct, in agreement with the green chemistry principles [13–16].
Among the catalysts active in the carbonylation reaction, the most used are those based on the palladium (Pd) complexes, which usually required the presence of phosphine ligands to lead to excellent conversions [17–22]. The use of phosphines, however, is undesirable because of their toxicity and air as well as moisture sensitive with conversion to, for example, phosphine oxide species.
In the present paper the authors propose a phosphine-free alkoxycarbonylation of iodobenzene, catalyzed by a heterogeneous Pd metal catalyst, to produce esters of the benzoic acid (Scheme 1).
From an industrial point of view, although the activities of homogeneous Pd-catalysts are usually very high, the treatments necessary to recover the precious metal (Pd) may be very costly. As matter of fact, the recovery and reuse of catalysts are the two most important features for many catalytic processes . Therefore, heterogeneous catalysis appears particularly attractive as it allows the production and ready separation of large quantities of products.
To further reduce the costs of process, recently many researchers are proposing the use of magnetically recoverable catalysts (MRCs) as an interesting alternative to the filtration or centrifugation procedures to separate the solid catalyst from the liquid products . Effectively, MRCs, which need of an external magnet, nowadays have found wide applications in various areas of industrial production [29–32]. In particular, heterogeneous catalysts consisting of palladium nanoparticles supported on magnetic Fe3O4 (magnetite) have found applications in several important reactions [33–38].
We propose the synthesis of new magnetically recoverable Pd-metal catalysts, active in the alkoxycarbonylation of iodobenzene without addition of phosphine ligands. The new catalysts have been synthesized by deposition of Pd-metal on inert supporting materials containing a percent of magnetite. The magnetite allows the complete recover of the catalyst with the simple application of an external magnetic field.
Carbon monoxide and ethene were supplied by SIAD Company Italy (‘research grade’, purity >99.9%); iodobenzene, triethylamine (NEt3), methanol, ethanol, propanol, 1-butanol, 2-butanol, phenol, FeCl3·6H2O, FeCl2·4H2O, NaOH, sodium dodecylsulfate (SDS) were purchased from Sigma-Aldrich and used without further purifications.
Equipments and characterization
The catalyst precursors were weighted on a Sartorious Micro balance (precision 0.001 mg).
Gas-chromatographic (GC) analysis was performed on a Hewlett Packard Model 7890, Series II chromatograph fitted with a HP5, 30 m × 0.35 mm × 0.53 μm column (detector: FID; carrier gas: N2, 0.2 mL/min; oven: 318 K (3 min) to 523 K at 15 K/min).
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 °C/min).
Fourier transform infrared (FTIR) spectra were recorded on a Nicolet Magna 750 instrument in KBr powder.
The 1H and 13C NMR spectra were recorded on a Bruker Avance 300 MHz spectrometer. Chemical shifts (δ) are reported in ppm, using TMS (δ=0) as an internal standard in 1,1,1,3,3,3-hexafluoro-2-propanol/CDCl3:10/1 mixture.
The inductively coupled plasma optical emission spectrometry (ICP-OES) analyses, to identification and detection of trace metals, were performed by using the PerkinElmer® Optima 7300 DV ICP-OES instrument.
The specific surface area was measured by using the Micromeritics Instrument Inc. (USA) apparatus, ASAP 2010 model.
Synthesis of magnetite (Fe3O4)
Magnetite nanoparticles (Fe3O4) have been synthesized as reported in the literature . To avoid the formation of both maghemite, γ-Fe2O3, and hematite, α-Fe2O3, the synthesis was performed under oxygen-free conditions.
In a typical experiment, 0.01 mol FeCl2·4H2O and 0.02 mol FeCl3·6H2O were dissolved in 100 mL degassed deionizer water, mechanically stirred at 500 rpm while nitrogen gas flowed into the flask. Under vigorous stirring, 100 mL of a NaOH 1 m solution was added to the Fe2+/Fe3+ solution, drop by drop (in 10 min), and the stirring continued for 3 h. After this reaction time, the solution was decanted, allowing the black particles to be washed with degassed and deionized water. This procedure was repeated three times, and then the particles were separated by a permanent magnet and dried in a vacuum desiccator.
The specific surface area of the powders was measured by the Brunauer–Emmett–Teller (BET) method and the pore size distribution of the powders by the adsorption–desorption isotherm utilizing nitrogen as the adsorbate after drying at 353 K for 12 h.
Polyketone (PK) synthesis
The perfectly alternated polyketone (poly(1-oxo-trimethylene), PK) has been used as inert solid to support the metal catalyst. The polymer has been synthesized as reported in the literature from the catalyzed reaction between CO and ethene (see Scheme 2) [39, 41].
In a typical experiment, 1.00 mg of [Pd(OAc)2(dppp)] (1.57 10−3 mmol) was added to 80 mL of solvent (water containing SDS, 20 mm) placed in a Hastelloy C autoclave of ca. 250 mL provided with a four-blade self-aspirating turbine. The reactor was flushed with a 1/1 mixture of CO/C2H4 at room temperature with stirring and then pressurized with 0.5 MPa of the gas mixture. The reaction mixture was heated to 363 K in ca. 10 min without stirring and the pressure was then adjusted to the desired value (typically 4.5 MPa total pressure). The temperature and the pressure were maintained constant throughout the experiment (1 h, rate stirring 700 rpm). At the end of the experiment the reactor was quickly cooled and carefully depressurized. The polymer was completely precipitated by addition of 100 mL of acetone and the slurry obtained was filtered, washed with water and acetone and dried under vacuum at 343 K. The dried polymer was weighed and characterized by the IR and NMR spectroscopies.
The FT-IR spectra show typical stretching signals of CO groups at 1695 cm−1 and –CH2– groups at 2915 cm−1. The 13C NMR spectra, shows a single carbonyl absorption at 212.65 ppm (–C (O)CH2CH2–) and a single resonance for the CH2 groups at 35.73 ppm (C(O)C H2C H2) in the ratio 1:2 due to the exclusive perfectly alternated structure .
Synthesis of Fe3O4/PK magnetic support
The synthesis of magnetic supports has been performed by incorporating Fe3O4 nanoparticles on PK (named aFe3O4/PK). In a typical procedure, 0.200 g of Fe3O4 suspended in 60 mL of water containing SDS (sodium dodecylsulfate, 20 mm) as surfactant, was used as a solvent for the emulsion CO–ethene copolymerization in the presence of dichloromethane (20 mL) . The reaction solution was mechanically stirred at 600 rpm for 1–2 h. After this reaction time, the reactor was cooled and depressurized. Noteworthy, it has been carried out more experiments at different reaction times in order to obtained inside the reactor the 15% aFe3O4/PK support, used for the comparative experiments. The reaction solution was decanted, allowing the particles to be washed with degassed and deionized water. The Fe3O4/PK particles were separated by a permanent magnet and dried in a vacuum desiccator. The amount Fe3O4 deposed on the support was determined through the ICP analysis (ca. 15%) and the specific surface area was measured by using the fisi-sorption-BET technique (Table 1).
Alternatively, the magnetic support, 15% Fe3O4/PK, has been synthesized by deposition of Fe3O4 nanoparticles on the preformed PK following the same procedure adopted for the synthesis of Fe3O4/C (named bFe3O4/PK).
Synthesis of Fe3O4/C magnetic support
The synthesis of magnetic supports Fe3O4/C (ca. 15%) has been performed by deposition of Fe3O4 nanoparticles on carbon surface. In a typical procedure, 0.150 g of Fe3O4 suspended in 5 mL of methanol was added under vigorous stirring to 1 g of carbon in 20 mL of methanol. The reaction solution was mechanically stirred at 500 rpm and refluxed for 2 h. After this reaction time, the solution was decanted, allowing the particles to be washed with degassed and deionized water. The Fe3O4/C particles were separated by a permanent magnet and dried in a vacuum desiccator. The amount Fe3O4 deposed on the support was ca. 15% (determined through the ICP analysis) and the specific surface area was measured by using the fisi-sorption-BET technique (Table 1).
Synthesis Pd-metal supported catalyst
The Pd metal catalyst was successfully supported on Fe3O4/PK (but also on Fe3O4/C, Fe3O4, PK or C) by using a facile metal adsorption-reduction procedure. In a typical procedure, 1.000 g of magnetic support (with ca. 15% of Fe3O4) was added to a solution formed by 20 mg of Pd(OAc)2 [Pd(II) acetate] in 20 mL of methanol under vigorous stirring. The suspension was mechanically stirred at 500 rpm and refluxed for 1–2 h. After this reaction time the reduction of Pd(II) to Pd metal is completed and the Pd/Fe3O4/C (but also the Fe3O4/PK, Fe3O4, PK or C) magnetic catalyst formed was separated by a permanent magnet, washed and dried in a vacuum desiccator.
The amount Pd-metal deposed on the support was determined through the ICP analysis (in all the catalysts was ca. 1% respect to the support) and the specific surface area was determined by using the fisi-sorption-BET technique.
The catalysis was carried out in a stainless steel batch reactor of ca. 100 mL. In a typical experiment, 50 mg of Pd-based catalyst, together with 4.49 10–3 mol of iodobenzene (915 mg) and 5.25 10–3 mol of triethylamine (NEt3), were added to 20 mL of alcoholic solvent (for instance MeOH).
In order to avoid catalyst deactivation, due to the presence of oxygen, the reactor was carefully flushed with CO at room temperature with stirring and then pressurized with 0.5 MPa of CO and heated up to 393 K in ca. 10 min without stirring. The pressure was adjusted to the desired value (typically 4.5 MPa total pressure) and, while stirring, maintained constant throughout the experiment (1–4 h) by continuously supplying the carbon monoxide from a reservoir.
At the end of each experiment the reactor was quickly cooled and carefully depressurized. The separation of the solid catalyst by the liquid mixture has been carried out by using a strong permanent magnet in order to ensure that almost no particles of Pd metal were lost.
The reaction products were detected and quantified by the GC and GC-MS analysis.
The recycling experiments need the quantitative recover of the magnetic catalyst. The solid has been washed with methanol for more times and dried in a vacuum desiccator.
As the catalyst has been recovered, washed and weighted, it has been reused in the successive reaction. After refilling with fresh reaction solution, the reaction with the recycled catalyst has been carried out following the same procedures above described.
Results and discussion
On the magnetically recoverable catalysts
The MRCs have been obtained by deposition of Pd-metal on supports containing magnetite which allows the magnetic recovery. The catalysts differ by the nature of the support which are formed by the perfectly alternated CO–ethene copolymer poly(1-oxo-trimethylene) (PK) containing magnetite (Fe3O4/PK), magnetite on carbon (Fe3O4/C), and pure magnetite (Fe3O4). The PK is a versatile polymer having high chemical resistance towards several common solvents, which makes it suitable as supporting material for catalysts. In the present paper two types of Fe3O4/PK supports have been prepared which differ in the mode of Fe3O4 deposition (see experimental). The copolymerization carried out in a water emulsion, containing magnetite nanoparticles, leads to a polymer in which Fe3O4 remain englobed. Alternatively the deposition of Fe3O4 on PK, as well as on carbon, has been obtained following a simple impregnation procedure.
The Table 1 shows the specific surface area measurements performed on the supports and on the Pd-catalysts.
The Fe3O4/C support showed a specific surface area lower than the pure carbon, whereas the 15% magnetite on PK slightly increases the specific surface area respect to the pure polymer. Such results is probably due to the mixing of ca. 15% of magnetite, having a specific surface area (75.1 m2/g) lower than carbon (980.6 m2/g) but higher than the pure PK (13.6 m2/g).
As expected, the deposition of ca. 1% Pd-metal on the surface of each supporting material slightly decreases the specific surface areas, as shown in the Table 1.
The methoxycarbonylation of iodobenzene
The conversions obtained with the magnetic (Pd/Fe3O4/C, Pd/aFe3O4/PK and Pd/bFe3O4/PK) and non-magnetic (Pd/C and Pd/PK) catalysts are showed in the Figure 1. All the experiments lead to iodobenzene conversion higher than 80%, with 100% selectivity.
The comparative analysis of such results suggests that the deposition of 15% of Fe3O4 on the surface of support has a small influence on the catalyst performance.
Furthermore, the order of catalytic activity Pd/Fe3O4/C=Pd/C>Pd/Fe3O4>Pd/Fe3O4/PK=Pd/PK suggests a correlation with the specific surface areas of the supporting materials (C>Fe3O4>PK, Table 1).
The influence of batch time on conversion is showed in the Figure 2: under the reaction conditions specified, all the catalysts studied reach 100% of iodobenzene conversion (100% selectivity) in reaction times lower than 3 h.
It is noteworthy that in all the experiments the Fe3O4, deposited on different supporting material, preserved its magnetic properties during the reaction cycles. At the end of each reaction the catalyst was readily recovered by using an external magnet and successfully reused in other reactions.
Stability tests of the Pd/aFe3O4/PK catalyst have been carried out as function of pH (range of pH=3–10 obtained through addition of HCl or NaOH) in methanol suspension. The ICP measurements showed that after 4 days of exposure at neutral conditions about 1.0% of the iron was dissolved, lost from the solid phase. When basic conditions were chosen, even less iron was dissolved (e.g. at pH=8.5 only 0.04% of the total iron). Naturally, lower pH values led to higher dissolution rates. At pH=5–3 about 3%–4.5% of the total iron was dissolved after 4 days, respectively. In the same tests by using Pd/Fe3O4/C or Pd/bFe3O4/PK catalysts suspended in methanol at neutral conditions, about 1.7% of the total iron was dissolved, whereas at pH=8.5 about 0.1% and at pH=5–3 about 6%–7.5%, respectively. As the dissolution of iron is correlated with the dissolution of the magnetite, the results suggest a slightly better stability of the magnetite englobed into the PK during the copolymerization respect to the magnetite deposited on the preformed PK or on carbon.
As the recycle of the catalyst is essential for an economically feasible application, we have been carried out five experiments by reusing the same catalyst. Among the catalysts used, the Pd/Fe3O4, Pd/Fe3O4/C and Pd/bFe3O4/PK have lost the magnetic properties after three or four recycles, whereas the Pd/aFe3O4/PK was magnetically recovered also at the end of the 5th recycle.
Although leaching of Fe and Pd has been measured (by ICP measurements) for all the catalysts, those in which magnetite is incorporated into the bulk of the poly(1-oxo-trimethylene) (Pd/aFe3O4/PK) showed a lower loss of Fe together with a higher stability of the magnetic properties. In such a catalyst, the deactivation of magnetic properties is less probable, being magnetite protected by the polymer itself. However, by increasing the number of recycles the conversion slightly decreases, remaining in any case higher than 80% at the end of the 5th recycle. The ICP measurements at the end of such recycles showed that the leaching of Pd was in any case less than 5% of the initial Pd metal.
The performance of the Pd/aFe3O4/PK catalyst has been further studied. The Figure 3 shows the conversion as a function of both temperature and reaction time. By increasing the temperature up to 423 K the 100% of conversion is obtained in about 2 h without catalyst deactivation. The apparent activation energy evaluated by the Arrhenius plot was about 10.8 kcal/mol suggesting a reaction rate not influenced by diffusion phenomena.
The influence of the nature of alcohols in the alkoxycarbonylation of iodobenzene
The influence of the nature of alcohols has been evaluated under the condition of the Figure 4. The conversion decreases following the increase of pKa values of alcohols and the increase of steric hindrance in the order CH3OH<C2H5OH<1-C3H7OH<1-C4H9OH<2-C4H9OH. Such results accord with the studies reported in literature [43, 44] on the order of alcohols reactivity in the reaction with acyl–palladium complexes.
The magnetic Pd/aFe3O4/PK and Pd/Fe3O4/C catalysts lead to comparable conversions in each alcoholic solvent tested, suggesting a negligible influence of the support on the catalytic activity.
According to the above discussion, the highest conversion has been obtained by using phenol as a solvent. The conversion to phenyl benzoate is higher than 97% with both Pd/aFe3O4/PK and Pd/Fe3O4/C catalysts.
The influence of different solvents on the reactivity of phenol has been also evaluated. Although in all the experiments the conversion obtained was lower than in pure phenol, mainly due to its lower concentration, it was possible to point out some qualitative differences, depending by the nature of the solvent. Figure 5 shows that the conversion increases by increasing the dielectric constant of the solvent [ε=cyclohexane 2.02, toluene 2.38, tetrahydrofuran (THF) 7.58, nitromethane 35.87, dimethyl sulfoxide (DMSO) 46.7], suggesting the influence of a solvent effect on the stability of ionic species (for instance phenolate), active in the rate determining step to form the ester. Furthermore the more polar solvents show a better surface affinity to PK respect to carbon, whereas it is the contrary for the more apolar solvents.
Such results, together with the different influence of alcohols (Figure 4), further suggest a possible mechanism, which involves the outer-sphere attack of the alcohol (or alkoxide) at the acyl carbon atom [43, 44].
In a plausible reaction mechanism the iodobenzene readily adds to Pd metal [or Pd(0) species on the surface, step 1 in the Scheme 3] forming a Pd(II)-phenyl species which inserts carbon monoxide (step 2). The Pd-acyl species formed reacts with phenol or alcohols (methanol in the Scheme 3, step 3) to lead the corresponding ester. The reductive elimination of HI from the Pd(II) intermediates, favored by the presence of the base (NEt3), regenerates the Pd-metal (or Pd(0) species) catalyst on the surface of the support (step 4), which can restart the cycle. It is not clarify how such possible Pd(II)-intermediates remain on the surface (dotted line in the Scheme 3). Some hypothesis are plausible: for instance the Pd(II) species are linked to the Fe3O4 , or alternatively they are in the solvent inside the pores (homogeneous catalysis) where however they are getting trapped. Similarly to the classical synthesis of the Pd-supported catalyst (see Experimental), the reductive elimination of HI in the step 4 forms the Pd metal which is again deposed on the catalyst surface.
In the present paper, such point has been not further investigated, but the hypotheses reported accord with the low Pd-leaching and the high activity obtained by using the recycled catalysts.
We have been tested some magnetically recoverable Pd-metal catalyst in the alkoxycarbonylation of iodobenzene under phosphine-free conditions. Very interesting results have been obtained by using as a new supporting material the polymer poly(1-oxo-trimethylene) containing ca. 15% of magnetite incorporated in the bulk. The conversion to benzoate esters has been optimized and compared to a magnetically recoverable Pd/Fe3O4/C and Pd/Fe3O4 catalysts. At 393 K, the new catalyst tested has been reached a conversion of 100%, in a reaction time lower than 3 h. Such catalyst has been completely recovered with an external magnet and efficiently recycled for more than five times. The catalytic activity slightly decreases by increasing the number of recycles, remaining in any case higher than 80% at the end of the 5th recycle. Although leaching of Fe and Pd has been measured for all the catalysts tested, by incorporating magnetite into the bulk of the poly(1-oxo-trimethylene) (Pd/aFe3O4/PK), the loss of Fe3O4 is lower than in the other catalysts, probably because the magnetic particles are protected by the polymer chains.
Financial support by Ca’ Foscari University of Venice is gratefully acknowledged (ADIR fund 2013–2014).
E. I. Negishi. Handbook of Organopalladium Chemistry for Organic Synthesis, Wiley, New York (2002). Google Scholar
A. de Meijere, F. Diederich (Eds.). Metal-catalysed Cross-Coupling Reactions, 2nd ed. Wiley-VCH, Weinheim (2004).Google Scholar
M. Beller, C. Bolm (Eds.). Transition Metals for Organic Synthesis, 2nd ed., Wiley-VCH, Weinheim (2004). Google Scholar
H. M. Colquhoun, D. J. Thompson, M. V. Twigg. Carbonylation Direct Synthesis of Carbonyl Compounds, Plenum Press, New York (1991).Google Scholar
Y. V. Gulevich, N. A. Bumagin, I. P. Beletskaya. Russ. Chem. Rev. 37, 299 (1988).Google Scholar
J. Tsuji. Palladium Reagents and Catalysts, pp. 265, Wiley, Chichester, UK (2004). Google Scholar
M. Mori. in Handbook of Organopalladium Chemistry for Organic Synthesis, vol. 2, E. Negishi (Ed.), pp. 2663, Wiley-Interscience, New York (2002).Google Scholar
M. Benaglia. Recoverable Organic Catalysts, in Recoverable and Recyclable Catalysts, M. Benaglia (Ed.), John Wiley & Sons, Ltd, Chichester, UK (2009). Google Scholar
M. C. Mascolo, Y. Pei, T. A. Ring. Materials. 6, 5549 (2013). Google Scholar
P. W. N. M. van Leeuwen, Homogeneous Catalysis Understanding the Art, Kluwer Academic Pub, Dordrecht, The Netherlands (2004). Google Scholar
About the article
Published Online: 2016-06-10
Published in Print: 2016-05-01
Citation Information: Pure and Applied Chemistry, Volume 88, Issue 5, Pages 445–455, ISSN (Online) 1365-3075, ISSN (Print) 0033-4545, DOI: https://doi.org/10.1515/pac-2016-0204.http://creativecommons.org/licenses/by-nc-nd/4.0/.