Definition of green synthetic tools based on safer reaction media, heterogeneous catalysis, and flow technology

Introduction

Green (Sustainable) Chemistry can be nowadays considered a scientific platform where academic and industrial researchers focus their efforts towards the optimization of a chemical process both from chemical and environmental point of views. With all its multifaceted interests and applications Green Chemistry is the arena where scientists from several different areas can cooperate and offer a multidisciplinary approach to the solve the modern issues related to a sustainable development.

Accordingly, in the last 20 years the scientific community has gained an increasing familiarity with the “Green Chemistry” principles as witnessed by the growing number of scientific contributions from several different areas of research.

Due to the multidisciplinarity of Green Chemistry, additional specifications are often needed to better frame the specific contributions in this wide research area.

In our case, we have had a long-term interest in defining novel protocols aimed at minimizing the waste associated to chemical synthesis. Special attention has been devoted to the use of solvent-free conditions (SolFC) or safe and recoverable reaction media, recoverable catalysts featuring a minimal leaching of the active organometallic or organic species, and setting of flow reactors to maximize the chemical efficiency and simplify the reuse of the catalytic system and the recovery of the solvent if used.

In this review article our approach towards the definition of sustainable synthetic protocols will be presented. Among the many different possible approaches, in the last few years we have been directing our efforts towards the development of three different areas of research that we believe are of fundamental importance and may serve as key general tools for developing greener organic chemistry methodologies: (i) safer solvents/reaction media, (ii) heterogeneous catalysis in greener media, (iii) flow technology. The results obtained are discussed below.

Solvent/reaction medium

In the production of complex molecules such as in fine chemicals and pharmaceuticals synthesis, all the solvents used to run a reaction and to isolate and/or purify the products, represent the major component in terms of mass and volume of the entire process [1]. It is therefore evident how crucial is the choice of the solvent used to minimize the environmental footprint of the chemical production [2], [3].

It is therefore obvious that major attention must be devoted to the selection of the most adequate solvents to achieve the highest chemical efficiency but also to avoid the use of most problematic and widely used traditional reaction media that derive from non-renewable fossil resources and that features high toxicity and bio-accumulation potential [4], [5], [6], [7], [8], [9].

Several promising directions have been undertaken by the green chemistry community and excellent results have been achieved using non-conventional solvents as water [10], [11], [12], [13], [14], [15], supercritical CO₂ [16], [17], [18], [19], ionic liquids [20], [21], [22] as well as biomass-derived solvents that feature the obvious advantage of deriving from renewable resources [23], [24], [25]. In the latter case, additional considerations should be made by evaluating the synthetic route used for their preparation but can be certainly considered as “greener” and promising alternatives to fossil based solvents.

Our research is mainly focused on the use of water, solvent-free conditions (SolFC) and biomass-derived solvents. We are also investigating the use of aqueous azeotropes as recoverable alternative to the commonly used water/organic solvent mixtures that being not easily recoverable represent a significant source of waste.

We have recently focused our main interest towards polar aprotic solvents that are of major concern in terms of safety and toxicity [6] but that are also extremely useful for performing modern synthetic transformations specifically in the case of cross-coupling reactions and C–H functionalizations.

In our search for novel solvent candidates deriving from biomass, we have recently proposed γ-valerolactone (GVL) as a possible replacement for common polar aprotic solvents as its polarity is comparable to representative DMF or N-methylpyrrolidone (NMP) (ε=36.5, 36.7 and 32.0, respectively). Therefore, we have investigated the use of GVL in those chemical processes that more often rely on the use of polar aprotic media, i.e. cross-coupling reactions.

GVL was first prepared in 2004 by Horvath et al. [26], [27], [28] and is nowadays prepared by hydrogenative cyclization of levulinic acid, a platform chemical derived from lignocellulosic biomass.

Our first application of GVL as reaction medium was in the Mizoroki-Heck reaction catalyzed by commercially available Pd/C (0.1 mol%) [29]. Iodoarenes 1 and styrenes or acrylates 2 reacted in the presence of 1 equiv. of triethylamine always leading to the products in very satisfactorily yields (80–90%) (Scheme 1).

Scheme 1:

Mizoroki–Heck reaction catalyzed by Pd/C in GVL as reaction medium.

At the end of the process the solid catalyst was filtered off and proved to be reusable. Afterwards, addition of water, that is completely miscible with GVL, induced the precipitation of the product 3 that could be isolated by simple filtration as pure compound. This protocol proved to be much more efficient in terms of waste produced compared to typical procedures requiring aqueous work-up.

The study was completed by comparing the results obtained in GVL with those yielded in other common reaction media (DMF, NMP, DMSO, acetonitrile, water and a 4:1 mixture of water and DMF) and besides the comparable reactivity, GVL showed to be superior in terms of the amount of metal leached into the product. In fact, inductively coupled plasma-atomic emission spectrometric (ICP-AES) analyses revealed how strongly dependent is the metal leaching by the medium used in the process. In fact, while in the reaction with GVL only 3.9 ppm of palladium were found in the product, 50 and 279 ppm were, respectively detected in DMF and NMP. This result may be attributed to the lower ability of GVL to form stable complexes with soluble palladium species formed during the usual outcome of a cross-coupling reaction. This result is crucially important especially in terms of waste production, because the amount of metals allowed in pharmaceutical compounds is strictly regulated, and purification from metal is possible but costly [30], [31], [32].

In certain cases, products cannot be easily further purified and the amount of metal leached into the product is dictated by just the protocol used in the cross-coupling reaction. As an example we have reported the synthesis of a polymeric organic semiconductors (poly(2,5-dihexylphenylenedivinylene-alt-1,4-phenylenevinylene) 6, (Scheme 2) whose performance of the corresponding device can be negatively affected by the presence of metal species [33], [34].

Scheme 2:

Synthesis of a model semiconductor through a Mizoroki–Heck reaction catalyzed by Pd/C in GVL.

Semiconductor 6 was prepared by a Pd/C-catalyzed Mizoroki–Heck reaction in GVL or NMP (Scheme 2). While product yield was similar in the two solvents (~70%), the use of GVL led to a dramatically lower metal contamination of the product compared to NMP (6 ppm vs. 860 ppm).

With a similar approach we have showed that similar results can be obtained by using GVL as medium for Pd/C catalyzed Sonogashira-Hagihara reaction [35] and in the the Hiyama reaction [36].

The different ability of GVL to dissolve palladium from the heterogeneous support, was proved again observing that in GVL the Sonogashira coupling is slower but led to significantly lower metal contamination (19 ppm compared to 144 ppm in NMP, 384 ppm in DMSO and 617 ppm in DMF).

More recently, we have also showed that GVL is a promising medium for processes involving C–H activations and using this medium we reported the first example of a Catellani reaction catalyzed by a heterogeneous palladium catalyst [37]. Mechanistic investigations led to the conclusion that in the case of Pd/Al₂O₃, this catalyst likely operated via a genuinely heterogeneous mechanism and could consecutively reused four times without change in its efficiency. In addition, the very small palladium-leaching observed in solution (2.2 ppm) further supported the longevity of the solid catalyst under our conditions.

GVL has proved to be an effective medium also for the C–H arylation of 1,2,3-triazoles catalyzed by commercially available Pd/C (5 mol%) [38]. Besides the large scope of the process, mechanistic studies based on hot-filtration, mercury poisoning and three-phase tests, provided support for a heterogeneous catalytic mechanism and the amount of palladium leached in solution was found to be always below 5.5 ppm.

As a continuation of our search for novel media for waste-minimized protocols, we have been paying attention also to the use of aqueous azeotropes. In fact, several common protocols in chemical synthesis employs homogeneous mixtures of water/organic solvent (such as EtOH, MeOH, but very commonly the toxic DMF and THF) where their composition is dictated more by the traditional experience than by a precise need or optimization. Consequently, in most of the cases these aqueous mixtures cannot be easily recovered resulting in the significant production of a costly waste.

We have recently reported the use of acetonitrile azeotrope in several cross-coupling reactions [39], [40], [41] but most recently we have focused our attention on waste minimized protocols for copper-catalyzed azide alkyne cycloaddition (CuAAC) a typical click reaction [42], [43], widely applied in countless areas for the preparation of target materials. For recent representative examples see [44], [45], [46] and literature cited herein.

CuAAC is a typical example of a process involving organic molecules (azides, alkynes and 1,2,3-triazoles) highly insoluble in water and inorganic reagents or catalysts insoluble in organic solvents. In fact, salts such as CuSO₄ or Cu(OAc)₂ and sodium ascorbate or ascorbic acid used generating in situ the actual Cu(I) catalytic species. This process is therefore generally performed in aqueous media while the organic co-solvent is generally an alcohol (t-BuOH, EtOH, MeOH), or DMSO, acetonitrile, THF or CH₂Cl₂.

In the greenest protocols, final products are isolated by simple re-crystallization after adding more water to induce precipitation.

We have directed our attention towards furfuryl alcohol (FA) that is a clear liquid that forms an azeotrope with water (20 wt% of FA) with a boiling point of 98.5 °C. Interestingly, this alcoholic azeotrope is formed with large amounts of water and therefore if used in CuAAC can induce directly the precipitation of the final triazole product without the need for adding and waste more water.

In this perspective, we hypothesized that the use of FA/water azeotrope could be effective in the definition of a waste minimized protocol for CuAAC provided that this azeotropic mixture could be efficiently recovered and reused minimizing its release in the environment. FA/water azeotrope proved to be widely efficient in this process allowing to isolate the expected triazoles in good to excellent yields and often in shorter times compared to the reactions conducted in reference t-BuOH/water solvent mixture. FA/water azeotrope is easily recoverable by simple distillation and our procedure allowed a significant reduction of waste produced if compared to classical CuAAC procedures, as confirmed by the comparison of the E-factor values (4.3 vs. >100) [47].

Our group is further investigating the use of novel solvent mixtures to reduce the waste of synthetic procedures as well as the search for innovative chemical outcomes to innovate fundamental and useful synthetic transformations.

Heterogeneous catalysis in greener media

The use of catalysis for the activation of inert bonds (or reactants) plays a pivotal role in the definition of effective synthetic methodologies fulfilling the key principles of Green Chemistry. Among the different approaches for the definition of catalytic methodologies, the use of insoluble catalysts that operate in a heterogeneous fashion is highly intriguing and also challenging [48], [49], [50], [51], [52], [53]. Significant advantages may derive from the adoption of this strategy with the reduction of waste produced, provided that an easy recovery and reuse of the catalyst is achieved. Moreover, in organometallic catalysis the use of a heterogeneous catalytic system can result in a reduced metal contamination of the desired products, which is crucially important in various applied areas.

A wide variety of different heterogeneous supports have been used to immobilize a catalyst going from organic (polystyrenes, polymethacrylates, etc) to inorganic materials (metal oxides, silica, zirconia, etc), as well as hybrid materials (e.g. metal organic frameworks-MOF).

The most interesting results have been obtained when the solid supports not only possess the role of heterogenizing the catalyst but also play a relevant role in influencing the chemical efficiency and the selectivity of a process [54], [55], [56].

We are directing our efforts in the use of different supports such as silica [39], [40], [57], [58], zirconium phosphate or phosphonates [41], [59], [60], [61], and cross-linked polystyrenes [62], [63].

In all the cases the design of the support used for the preparation of the catalytic system should consider the reaction medium used for the process. The heterogeneous support should be able to allow the access of the reaction mixture to the catalytic site and generally, the solvent used as reaction medium plays a crucial role in allowing this process to occur efficiently.

For instance, widely used Merrifield resins consist in an insoluble polystyrene polymer cross-linked with divinylbenzene units, typically swelled in DMF or in CH₂Cl₂, both solvents are of major concern in the Green Chemistry arena. In addition, our final goal is to employ heterogeneous catalysts in flow reactors and a major issue that limits this application is the swelling properties of the resins that leads to a significant increase of their volume when treated with the proper solvent therefore requiring to operate a very high pressure [64].

We are currently focusing our attention towards the preparation of different cross-linked polystyrenes that differ from classic Merrifield resins for the cross-linker used. We have replaced divinylbenzene with larger molecules that not only could be able to enlarge the distance between the polystyrene chains but should be also able to confer different degrees of rigidity to the molecular structure to better tailoring the support to the specific solvent and reaction conditions and optimize the mobility of the chemicals at the catalytic sites (Fig. 1).

Fig. 1:

Novel “SPACE” heterogeneous polystyrenes as supports for catalytic systems.

A wide range of macroporous cross-linked copolymers of 4-vinylbenzylchloride (VBC) with divinylbenzene (DVB) to be employed as heterogeneous catalyst supports have been synthesized and characterized in terms of structure, thermal properties, and morphology.

The synthesis of such chloromethylated macroporous polystyrenic supports was successfully performed by employing different ratios of all reaction components, i.e. co-monomers, porogen, and stabilizer, as well as chemically different porogens to enable a broad variation in the (micro) structure of the materials.

In this regards, we also showed for the first time how Fourier transformed infrared spectroscopy mapping (μ-FTIR) may represent a powerful and easy-to-handle tool for the characterization of polymer supports as well as the corresponding immobilized catalytic systems, to examine functional groups distribution at a single-bead level. This is of particular interest since it helps in the prediction of resin behavior during chemical processes. Indeed, heterogeneous micro-domains within a bead may have important consequences for the macroscopic properties (e.g. swelling), and can lead to different reactivity of functional groups within a region of a bead.

Further, scanning electron microscopy (SEM), thermal analysis (MTDSC), and N₂ absorption (BET) for surface area and porosity measurements have been performed. Direct correlations between the polymerization conditions and resins chemical-physical properties have been observed [65], [66].

On the basis of the wide set of collected results, we identified three representative supports prepared by employing VBC/DVB=4/1, low molecular weight (13000–23000) polyvinylalcohol stabilizer (0.4%), and 2-ethylhexanoic acid (EEA), 1-chlorodecane (CD) or cyclohexanol (COX) (1:1 v/v) as porogens. The resin through EEA featured surface area (6.8 m²/g) and average pore volume (0.03 cm³/g) lower than those through CD (23 m²/g and 0.17 cm³/g, respectively) and COX (21 m²/g and 0.2 cm³/g, respectively). On the other hand, the resin through COX featured the highest average porous radius (398 Å vs. 140–160 Å).

Thus, such polystyrenic supports were used to immobilize a highly basic guanidine, i.e. 1,5,7-triazabicyclo[4.4.0]dec-5-ene (TBD) and the efficiency of the resulting polymeric TBD catalysts 7a–c was explored in the Michael reactions between 4-hydroxycoumarine 8 and a wide range of α,β-unsaturated ketones 9a–d to access Warfarin and its analogs 10a–d, which are important pharmacologically active molecules (Scheme 3).

Scheme 3:

Polymer supported base-catalyzed Michael addition of 4-hydroxycoumarin 8 to α,β-unsaturated ketones 9.

Good conversions were achieved (77–82%) for the reaction between 8 and 9a (equimolar ratio) when catalysts were used in 5 mol% amount at 70°C in cyclopentylmethylether (CPME), with 7a,c giving slightly better results. Notably, CPME has been selected since it is included in the list of green solvents [10].

To broaden the catalysts scope, ketones 9b–d were also employed in the Michael addition with 8, allowing the preparation of 10b–d in high to excellent yields. For all the substrates, the catalysts activity was 7b~7a>7c. These findings suggested that the supports prepared by using EEA and CD as porogens allowed to obtain the most accessible structure by substrates, despite their relative average pore size detected by BET analysis. This apparent discrepancy was tentatively explained taking into account that the average pore size may not reflect the pore size distribution.

Commercially available macroporous polystyrene based TBD was also investigated for comparison, and all the novel catalysts revealed to be superior to this latter in terms of catalytic activity. Finally, the catalysts were easily fully recovered by simple filtration and no loss of activity was observed after three cycles.

More recently [62], [63], we introduced a novel class of styrene-VBC gel-type resins to be employed as heterogeneous catalyst supports, containing the 1,4-bis(4-vinylphenoxy)benzene cross-linker (SPACeR, Fig. 1), which is a novel polar spacer larger than DVB. These polymers were designed to get increased compatibility with respect to sustainable reaction media, including solvent-free conditions, in organic synthesis.

Particularly, three novel chloromethylated polystyrene-based SPACeR (SP) resins featuring different networking loading capacities were successfully prepared by varying co-monomers ratio.

The resins were subsequently used to immobilize TBD and diethylamine, to access catalytic systems SP-TBD 11a–c and SP-TEA 12a–c (Scheme 4a), respectively, featuring different strength of the basic sites.

Scheme 4:

(a) Shorthand representation of catalysts 11–12; (b) Phenolysis of phenyl glycidyl ether 13 and phenols 14.

The supports as well as the relative catalysts were characterized in depth by elemental analysis, Fourier-transform infrared spectroscopy (FT-IR), thermogravimetric analysis (TGA), extensive solid state NMR experiments, scanning electron microscopy (SEM), and transmission electron microscopy (TEM). The efficiency of the catalytic systems was evaluated in a series of representative organic transformations, i.e. 1,2-epoxide ring opening, aldol-type condensation, and Michael addition, under solvent-free conditions. They were found to ensure a good catalytic efficiency which generally overcome that of commercially available macroporous polystyrene-supported TBD catalyst, thereby revealing themselves highly promising tools. An important aspect pointed out by these results is that the spacing between the linear polymer chains is large enough to enable easy access of the reactants to the active sites, without the help of a swelling medium.

Particularly, all the systems enabled optimal to excellent conversions (80–99%) for the reaction between phenyl glycidyl ether 13 and a range of phenols 14 (equimolar ratio) when catalysts were used in 5 mol% amount at 60°C (Scheme 4B).

As a general trend, a reduction of efficiency of the highly loaded 11a and 12a compared to the medium-loaded 11b and 12b, respectively, was observed, which may be likely due to crowding of the catalytic sites. Additionally, the low-loading catalysts 11c and 12c exhibits a decreased efficiency relative to the medium-loaded counterparts, which may be ascribed to a reduced mixing efficiency of the reaction mixture.

Not unexpectedly, the reactions mediated by SP-supported TBD 11a–c were faster compared to SP-TEA 12a–c, the active sites of these latter featuring a lower basic strength.

Finally, the catalysts could be easily recovered by simple filtration, and they substantially retained their activity up to three runs when reused.

Very recently [67] we developed new gel-type styrene-based resins containing the 1,4-bis(4-vinylphenoxy)benzene cross-linker (SPACeR, Fig. 1) to be used as supports to immobilize acid catalysts for the transformation of biomass-derived platform molecules into highly added value products.

The SPACeR resins were successfully prepared by suspension polymerization, and functionalized by sulfonation reaction to access solid acid catalysts 16a–c (Scheme 5). By varying the sulfonation conditions, systems featuring different networking loading capacities were easily afforded.

Scheme 5:

Catalyzed synthetic route to alkyl levulinates 19.

All the catalysts have been fully characterized in terms of structure, thermal properties, and morphology.

The efficiency of the catalytic systems has been evaluated in the esterification reaction of levulinic acid 17 with a range of alcohols 18 to access levulinates 19 (Scheme 5).

Levulinic acid has been recognized by the US Department of Energy as one of the top biomass-derived platform molecules, due to its chemical versatility and its relatively low cost-production from lignocellulose waste [68]. The catalytic upgrading of such important platform molecule into alkyl levulinate esters is currently of great interest. Indeed, levulinates are commonly employed as biofuel additives, flavoring agents, solvents, and as important intermediates of fine chemicals and bioplastics.

Interestingly, all the newly synthesized sulfonated polystyrenes revealed to be superior in terms of catalytic activity with respect to a range of ion-exchange resins available in the market, including Amberlyst 15^®, allowing the preparation levulinates in quantitative yields under mild reaction conditions. These findings open the route to the definition of an environmentally-friendly approach to the above target compounds.

Current research in this context is focused towards the preparation of macroreticular cross-linked polystyrenes and the definition of flow reactors for the preparation of levulinates under continuous-flow conditions.

Flow technology

In the development of a green synthetic protocol, much attention is generally paid to the selection of a proper reaction medium and the design of tailor-made heterogeneous/reusable catalytic systems, however the importance of the technology used for the stirring of the reaction mixture is not fully taken into account.

In order to minimize waste and reach the highest level of efficiency, minimal volumes of solvents and stoichiometric amounts of reagents should be mixed effectively through a solid catalytic system without compromising its morphology and structure and consequently facilitate its recovery and reuse. In fact, mechanical stirring inevitably causes the crunching of the solid catalytic system leading to its fine powdering that makes difficult its recovery and reuse.

In this context, the adoption of flow technology may be particularly effective allowing to mix small amounts of liquid through a reactor containing a solid catalyst without compromising its particle integrity and simplifying its recovery. In addition, flow chemistry represents a unique opportunity for innovating synthetic strategies and if properly applied, may be a powerful tool to minimize the waste associated to the chemical production of complex molecules [69], [70], [71].

We have used flow technology as an effective mixing technology in the combination of heterogeneous catalysts (organic and organometallic) and green reaction media such as azeotropes, water and also solvent-free conditions proving that with this approach the chemical efficiency as well as the greenness of a synthetic methodology can be effectively maximized and some examples will be described below [66], [69], [70], [71], [72], [73], [74], [75], [76], [77], [78].

A representative example of our approach on the use of flow conditions to minimize waste production has been disclosed for the Michael addition reactions of nitrocompounds to α,β-unsaturated compounds (Scheme 6) [72].

Scheme 6:

Preparation of γ-nitroketones 22.

By measuring the most relevant green metrics for the known protocols used to prepare γ-nitroketones, we have proven that adoption of the flow technology in combination with the use of a reusable heterogeneous catalyst, equimolar amounts of reagents and minimal amount of solvents, effectively leads to a highly chemically efficient procedure featuring a very low environmental impact (data calculations offered by Dr. John Andraos) [79].

The flow technology combined with the use of recoverable reaction media has allowed to develop sustainable methods for the heterogeneous palladium-catalyzed Suzuki–Miyaura, and Heck cross-coupling reactions [40], [41], [57], [61].

The use of flow approach has proven that many solid palladium sources operate via a “release and catch” mechanism [80] in which the active Pd species are released from the solid support during the oxidative addition step and then, at the conclusive elimination step when C–C bond is formed, recaptured onto the heterogeneous support. The effectiveness of this mechanism is related to the solvent used as medium, temperature and obviously to the capability of the heterogeneous support to stabilize palladium nanoparticles [81], [82], [83].

An early example from our group was the definition of an environmentally-friendly method for the Suzuki–Miyaura reaction using palladium supported on a cross-linked imidazolium network on silica gel (Scheme 7). The reaction conditions have been optimized by using an aqueous ethanol azeotrope, instead of classic not recoverable ethanol/water or DMF/water 1:1 mixtures as reaction media. This medium still furnishes the minimal amount of water that has a beneficial effect on the process but it is recoverable by distillation and therefore reusable minimizing the waste production. In addition, potassium carbonate was selected as the base needed in stoichiometric amounts to execute the process, due to its insolubility in ethanol/water azeotrope. The flow system was set by keeping the solid palladium catalyst separated from the base and also in these conditions high efficiency was achieved thanks to the “release and catch” mechanism. With this approach the solid catalytic system could be easily recovered and reused for four representative runs, with a TON of 3800. Also the products can be isolated by simple distillation of the reaction medium without the need of additional purification steps. The reduction of waste is impressive with an E-factor reduction from 3180–5100 to 3.5–3.9 compared to classic batch protocols that inevitably require purification steps and loss of the catalyst.

Scheme 7:

Waste-minimized, cyclic-flow protocol for the Suzuki coupling [57].

In many cases the adoption of SolFC may lead to more efficient and more selective reactions, with respect to the corresponding processes in solution, often also allowing the achievement of new chemical outcomes.

In addition, when heterogeneous catalytic systems are used, a lower efficiency is commonly observed if compared to non-supported counterparts and therefore promotion of an organic reaction becomes difficult. SolFC can be an effective approach to combine the need for waste minimized protocols while increasing the reactivity of the system, in order to allow the use of heterogeneous catalytic systems. Within this context the crucial role that the support may play under peculiar SolFC is evident and therefore the need for specifically designed supports adequate for their use in flow and under highly concentrated conditions or SolFC [62], [63], [65], [66], [69], [70], [78].

Silylated reactants represent a stabilized form of nucleophiles which can be easily and safely stored and that can be appropriately activated under catalyzed conditions to disclose their reactivity under adequately controlled conditions. For this reason, several silylated nuleophiles have been developed and originated a corresponding interesting chemistry [84], [85], [86], [87], [88], [89], [90].

Heterogeneous sources of fluoride, especially ammonium fluoride supported on polymers, have been used as nucleophilic catalyst for the activation of silylated nucleophiles, whilst more often their use is related to desilylation procedures [90], [91], [92].

The polystyrene supports typically used are either macroreticular or gel-type. The firsts are styrene-diviniylbenzene copolymers with a rather rigid structure and discrete pores, able to accommodate also large reactants. Gel-type resins, on the other hand, are characterized by a lower degree of cross-linking (typically <10%) and have no specific porosity.

The porous polystyryl-DABCOF₂ catalyst, a bis-ammonium fluoride salt of 1,4-diazabicyclo[2.2.2]octane supported on polystyrene resin, has been used to define a waste minimized protocol for the β-azidation of α,β-unsaturated ketones in water while the gel-type resin has been used to set up the preparation of 1,2-azidoalcohols by the activation of Si–N bond of TMS-N₃ nucleophile in SolFC [14], [93].

PS-DABCOF, the monoammonium fluoride derivatives of 1,4-diazabicyclo[2.2.2]octane supported on a gel type polystyrene, is a bifunctional fluoride source containing both a Lewis (fluoride) and a Brønsted (tertiary amine) basic site. It has been developed and applied under SolFC to promote the direct β-azidation of α,β-unsaturated carboxylic acid, without protection/deprotection steps of the acid functionality, for the preparation of β-azido and β-aminoacids [94], [95].

Commercially available Amberlyst fluoride (Amb-F) has been successfully used to set up one pot multi step protocols for the synthesis of β-hydroxyesters in flow by the Mukaiyama aldol addition of methyl trimethylsilyl dimethylketene acetal to a series of aldehydes based on the activation of Si–O bond by solid fluoride [96]. The activation of Si–N bond of TMS-N₃ nucleophile by Amberlyst fluoride has allowed the preparation of β-azido ketones and carboxylic acid in water and allowed to achieve multistep flow procedure in SolFC for the preparation of β-azido and N-Boc-β-amino ketones or N-Boc-γ-amino alcohols with high efficiency and very low E-factor [97], [98], [99].

Recently a novel protocol has been reported for the synthesis of β-cyanoketones from α,β-unsaturated ketones 26 by using TMSCN as a safer cyanide source and Amb-F as solid catalyst under SolFC [100]. A representative example is shown in Scheme 8. The recovered catalyst showed a decrease of the yield of the desired product 27, together with the formation of cyanohydrin trimethylsilyl ether deriving from the 1,2-addition reaction. The Authors speculated that in the recovered catalyst, the fluoride content was lower due to partial fluoride/cyanide anion exchange as later confirmed by elemental analysis and quantitative fluoride analysis by ionic exchange HPLC that showed a reduction of ca. 70%.

Scheme 8:

Amb-F catalyzed preparation of β-cyano ketones.

Since the recovery of the catalyst is fundamental for the efficiency and sustainability of a process, the Authors developed another protocol in two steps to obtain β-cyano ketones in high yields and low E-factors. A previous work [101] demonstrated that α,β-unsaturated ketones led to the corresponding cyanohydrin trimethylsilyl ethers when treated under solvent-free conditions in the presence of polystyryl-triphenyl posphine PS-TPP. This suggested the idea of a two-step flow process involving first a PS-TPP catalyzed formation of the cyanohydrin trimethylsilyl ethers, followed by an Amb-F promoted cyanide shift (1.2 to 1.4) to finally afford the desired β-cyano ketones (Scheme 9).

Scheme 9:

Two steps synthesis of β-cyano ketones in flow.

Solid acid catalyst have received much attention both for industry processes and academia since they are environmentally-friendly with respect the corrosiveness, safety, waste minimization and easy of recovery and reuse. Nevertheless little attention has been directed to the use of non-stoichiometric amounts of heterogeneous catalysts to define waste-minimized flow protocols.

The use of heterogeneous acidic catalysts in flow has been used from our group to define multistep waste-minimized protocols for the synthesis of β-hydroxy esters, and 1,2-azidoalcohols [93], [96].

Perfluorosulfonic acid (PFSA) resins are known as promising heterogeneous acid catalysts alternative to harmful mineral acids. Among them, Aquivion PFSA, a tetrafluoroethylene and perfluoro-2-(fluorosulfonylethoxy) vinyl ether copolymer, have a super acid character comparable to pure sulfuric acid [102]. These features could make Aquivion PFSA a suitable heterogenous catalyst for the synthesis of value added fine chemicals.

Recently, we have performed the synthesis of 2-pyrrolidin-2-ones, common molecular moiety found in biologically active products and pharmaceuticals [103], [104], by nitro-Mannich/Lactamization cascade reaction, employing catalytic amounts of Aquivion PSFA under SolFC or higly concentrated conditions in flow [105].

After optimization of the reaction conditions in classic batch and successful extension of the reaction scope to a variety of aromatic/aliphatic aldehydes and amines, a batch-like flow protocol was developed by packing the Aquivion PFSA as pellet formin a glass column. To avoid pump clogging caused by the viscosity of the reaction mixture, 1 mL/mmol of ethyl acetate was added and a 97% conversion to product 31 was achieved (Scheme 10). The solid acid catalyst was reused for other three consecutive runs, with high yield in all cases (91–94%).

Scheme 10:

Aquivion PFSA-catalyzed synthesis of 31 in flow.

To evaluate the greenness of the method, the component of the overall E-factor were calculated for both the batch and flow reactions. It appeared that E-aux (defining the waste associated to auxiliary materials used in the process as solvents for purification) represented the major contribution to E-total, but under flow conditions it was minimized, since column chromatography was avoided and the pure product was isolated by re-crystallization. As a result the flow protocol gives an E-total value of six without considering solvent recovery, which is lowered to ca. 2 when taking into account the recovery of the solvent, thus representing, respectively a 80 and 93% reduction with respect to the batch procedure (Table 1).

For comparison purposes, the authors calculated the same green metrics for some literature synthetic approach, and found E-total values greater than 300 in all cases.