Testing enabling techniques for olefin metathesis reactions of lipophilic substrates in water as a diluent

Summary Olefin metathesis reactions of diverse polyfunctional substrates were conducted in water emulsions using two hydrophobic ruthenium catalysts in the presence of air. Instead of using surfactants to increase the efficiency of the metathesis reaction in water, ultrasound and microwave techniques were tested on a small-scale reaction, whereas conventional heating and mechanical stirring were effective enough to provide high conversion and selectivity on a larger scale. The developed conditions extend known protocols for the aqueous metathesis methodology, utilizing relatively low catalyst loadings and allowing for simple product isolation and purification. The established synthetic protocol was successfully adopted in the large-scale synthesis of a pharmaceutically related product – sildenafil (Viagra) derivative.

Sustainable approach for metathesis reaction in water emulsion system on air.
Utilization of enabling techniques for boosting metathesis under aqueous conditions. RCM of medically important sildenafil derivative.

INTRODUCTION
Among the plethora methods for the construction of carbon-carbon double bonds, olefin metathesis has become one of the most convenient tool in chemists' hands (Grela, 2014;Ogba et al., 2018). This is undoubtedly associated with the development of new ruthenium catalysts with increased resistance to air and moisture. Similarly to other organometallic transformations, metathesis reactions are preferably carried out in carefully dried and degassed solvents from a petrochemical sources (benzene, toluene, and chlorinated solvents). However, owing to environmental concerns, making the olefin metathesis suitable for Green Chemistry guidelines remains a great interest in the scientific community (Clavier et al., 2007). Therefore, several methods were already developed to facilitate this transformation in the green solvents, including aqueous media (Burtscher and Grela, 2009;Zaman et al., 2009), to minimize the amount of waste (Kniese and Meier, 2010), and heavy metal residues in products (Szczepaniak et al., 2015;Vougioukalakis, 2012;Wheeler et al., 2016). In particular, olefin metathesis in water (or in its mixtures with polar solvents) represents an important solution for biocompatible synthesis, as well as reactions of highly polar substrates, which suffer from low reactivity in the traditional organic media (Lipshutz and Ghorai, 2014;Sabatino and Ward, 2019). Unfortunately, serious limitations for using aqueous media in metathesis reactions are associated with poor solubility of the commercially available ruthenium catalysts in water. To address this issue, several tailor-made water-soluble catalysts have been introduced, obtained by decorating classical ruthenium catalysts with ionic tags (Skowerski et al., 2012;Wang et al., 2015), and with other polar groups, such as PEGs (Hong and Grubbs, 2006). Recently, some of such catalysts have been commercialized (e.g. Ru8, Ru9Cl Figure 1) (Olszewski et al., 2020).
On the opposite scale of this problem are the water-insoluble (hydrophobic) starting materials, for which water may also be a potentially interesting medium during metathesis reactions. A lot of efforts were devoted to find more environmentally friendly conditions for the transformation of these ubiquitous hydrophobic substrates. In particular, it was found that carrying out reactions ''on water'', could be even more advantageous owing to the hydrophobic effect, which accelerates the reaction rate (Davis and Sinou, 2002). Nonetheless, reports on the use of common water insoluble metathesis catalysts under such heterogeneous aqueous conditions are still limited, probably owing to the problem of forming stable emulsions or micelles, facilitating the contact of catalyst with substrates. In a truly innovative approach, extensively explored by Lipshutz and co-workers, various amphiphilic additives (surfactants) allowed for performing the olefin metathesis reaction effectively ''in water' ' (Lipshutz et al., 2008;Lipshutz and Ghorai, 2010). In lack of reaction control, and increased risk of side product formation. In addition, in chemical transformations featuring solid starting materials or products, the difficulties in stirring the dense/pasty reaction mixture and problems in removing the product from the reactor shall be considered. In the particular case of olefin metathesis, faster decomposition of ruthenium catalyst under higher concentration can be expected sometimes, which might promote undesirable isomerization of substrates or products (Hong et al., 2005;Schmidt, 2004).
In this respect, it seems reasonable to us to use water as a diluent to suspend the reactants, ensuring convenient stirring and heat transfer. Moreover, in the context of waste management, water represents the major green solvent, since it is cheap, easily available, and non-toxic. However, the process of water purification is still expensive owing to relatively high heat capacity, which makes distillation processes energy consuming (Hä ckl and Kunz, 2018;Zhou et al., 2019). Therefore, the amount of water must be kept minimal whenever possible.
Herein we present our attempt for the olefin metathesis reactions under aqueous conditions, stressing the practical aspects of this processes, such as operational simplicity and efficient mass and heat transfer. These issues are crucial in performing reactions on a larger scale when one needs to consider not only the chemical but also the economical parameters of the transformation. To probe usefulness of the developed conditions on the manufacturing of fine chemicals and products of pharmaceutical interest, we decided to study a number of challenging lipophilic liquid and solid polyfunctional substrates in olefin metathesis using water as a reaction medium.

RESULTS AND DISSCUSION
Optimization study of olefin metathesis reaction in the aqueous emulsion system The development of ruthenium catalysts with increased stability and activity is one of the main trends in the advanced metathesis methodology. Peculiarly successful enhancement in activity has been achieved by introducing Hoveyda-Grubbs second-generation ruthenium complexes containing alkoxy styrene unit and N-heterocyclic carbenes (NHCs) as ligands, e.g. Ru1-Ru4 (Garber et al., 2000;Scholl et al., 1999). High functional group tolerance allows for its wide-spread applications in the synthesis of natural products, pharmaceuticals and target materials under mild conditions (Nicolaou et al., 2005). Recently, catalysts containing cyclic (alkyl)(amino)carbenes (CAACs) e.g. Ru5-Ru7 ( Figure 1) also had significant impact in further extension of the metathesis methodology, especially in the industrial production of fine chemicals (Gawin et al., 2017;Kajetanowicz et al., 2020;Lavallo et al., 2005;Marx et al., 2015). Nevertheless, in the area of aqueous metathesis, the ability to tolerate water and air by the ruthenium species is still one of the major limitations (Blanco et al., 2021;Guidone et al., 2015;Ton and Fogg, 2019).
To select the most robust and active system in the envisioned emulsion protocol, we tested a series of selected ruthenium catalysts Ru1-Ru9 in the ring closing metathesis (RCM) reaction of the standard model substrate, namely N,N-diallyltosylamide (1, DATA) floating under the water surface (Scheme 1). To ensure the best contact of the reactants, we used ultrasound that helped to create very small droplets of the organic phase and better stabilize the emulsion than mechanical stirring. The catalytic activity of Ru1-Ru9 was tested in non-degassed distilled water, under air in a standard laboratory ultrasound bath. More details concerning optimization studies are available in the supporting information (see Table S1), whereas the key results are summarized in Table 1.
Using commercially available catalysts Ru1 and Ru2, we observed rather moderate yields in the benchmark reaction under ultrasound-supported emulsification conditions (Table 1, entries 1 and 2). Respectively, catalyst Ru3 bearing a larger substituent in the NHC ligand showed much higher efficiency and gave the desired product in a 95% yield (Table 1, entry 3). This result is supported by higher stability and productivity generally observed for catalysts possessing a bulkier SIPr (1,3-bis-(2,6-triisopropylphenyl)-2-imidazolidinylidene) ligand in the structure (Clavier et al., 2009). Subsequently, we investigated the activity of catalyst Ru4 with an unsymmetrically substituted NHC ligand (uNHC) (Monsigny et al., 2021a) and, unfortunately, we noted low substrate conversion (Table 1, entry 4). In the next attempt, CAAC derivatives Ru5, Ru6, and Ru7 exhibiting relatively high protic solvent tolerance were tested (Blanco et al., 2021;Nagyhá zi et al., 2020). Owing to the superior activity of these ruthenium complexes connected with their slower initiation of metathetic cycle (Morvan et al., 2021a), we expected high activity in the RCM reaction of DATA (1). When Ru5 gave rather low conversion of starting material, catalyst Ru6 with electron withdrawing -NO 2 group in alkoxybenzylidene ligand performed the RCM reaction of 1 in a 77% yield (Table 1, entries 5 and 6). Even better results were noted for catalyst Ru7, which gave the corresponding RCM product 2 in a quantitative yield (Table 1, entry 7). As mentioned above, an important factor determining the utility of a given ruthenium catalyst for metathesis in water emulsion systems is the stability of the resulting propagating catalytic species (Blanco et al., 2021). For this reason, the next logical step was to check the catalytic capability of catalysts Ru8 and Ru9 tailored for application in aqueous metathesis (the latter was additionally altered for the metathesis reaction in aqueous conditions by tuning its counter ion, vide infra). Surprisingly, for the fully water-soluble complex Ru8, we observed a very low conversion of model substrate 1 (Table 1, entry 8).
We assumed that Cl À counterion of Ru8 is not lipophilic enough to provide good solubility of the catalyst in droplets of substrate 1; therefore catalyst remains in an aqueous layer, which results in significant decrease in its activity. The second reason stands behind the lower performance of Ru8 is unsuitable steric hindrance of the SIMes-type NHC ligand. In contrast, the polar but water-insoluble catalyst Ru9Cl bearing a bulkier SIPr-like NHC ligand led to a better result ( Table 1, entry 9). This encouraged us to further tune the lipophilicity of this catalyst. To do so, we exchanged the chloride counter ion for more lipophilic PF 6 À and BARF À ones (for the procedure, see Star Method section), and subjected the obtained catalysts Ru9PF 6 and Ru9BARF to the model metathesis reaction, observing increased productivity (Table 1, entries 10 and 11). Especially, the BARF complex (Ru9BARF) showed very high activity in the RCM reaction carried out in the emulsion system, giving the desired product in a 99% yield. Taking advantage of these results, for the most efficient complexes Ru7 and Ru9BARF, we lowered the catalyst loading in the same benchmark reaction and found that both catalysts gave excellent results even at loadings as low as 0.1 mol % in neat We postulated that the high performance of the aforementioned catalysts in the emulsion system is connected either with the presence of two (lipophilic) CAAC ligands or the SIPr-like NHC ligand together with BARF counter ion. Such structural properties provide a strong lipophilic character to these complexes. These features aid in the diffusion of the catalyst into emulsion droplets, which not only enhances the catalysts interaction with a lipophilic substrate but also prevents rapid decomposition of the catalyst by water. To the best of our knowledge, these examples represent an exceptionally efficient application of relatively low loading of ruthenium-based catalysts under aqueous metathesis conditions without the use of external additives or surfactants. Moreover, using this protocol, it was possible to avoid unwanted isomerization of the double bond in the product (caused by the decomposition of the ruthenium catalyst into Ru-hydride complexes and other species) (Bailey et al., 2017), which was previously observed in reactions carried out under classic conditions (Liu et al., 2014;Qiao et al., 2012), especially in polar solvents (Bantreil et al., 2012;Skowerski et al., 2014).

Scope and limitation study of aqueous emulsion olefin metathesis using ultrasounds
Encouraged by these results, we evaluated the most robust catalysts Ru7 and Ru9BARF in a variety of olefin metathesis reactions in neat water using acoustic emulsification, and the result of these studies is summarized in Table 2.
The established conditions were first evaluated in a series of RCM reactions with various dienes. We started our studies from classic substrate diethyl diallylmalonate (3) and observed that reaction proceed smoothly in neat water using only 0.5 mol % of catalyst Ru7 or Ru9BARF ( Table 2, entry 1). Similar result was noted for more demanding compound 5, when 1 mol % of Ru7 was enough to obtain a 90% yield of the desired product 6 ( Table 2, entry 2). Interestingly, the addition of a small amount of ethyl acetate as a co-solvent allowed us to achieved quantitative course of this transformation. An increased of the yield was especially visible in the case of Ru9BARF (from 41% in neat water to 95% with co-solvent). Whereas liquid and lipophilic substrates gave us gratifying results in RCM reactions leading to a five-membered ring, polar one was not reactive under examined conditions ( Table 2, entry 3). With these types of water-soluble compounds, the reaction could take place only in the region between an organic and aqueous phase because formed emulsion droplets contain only catalyst molecules. In this scenario, the probability of reaction between polar and highly solvated substrate and non-soluble in the water catalyst is dramatically lower. This fact also convinces us that emulsification of the reaction mixture is beneficial for rather lipophilic substrates, and the methodology is not suitable for highly polar substrates.
Next, we decided to focus on substrates known as more reluctant in RCM reactions. For this purpose, we choose dienes leading to products with trisubstituted C-C double bonds 9 and 11, and were pleased to obtain the cyclisation products 10 and 12 in very high yields in the presence of only 1 mol % Ru7 or Ru9BARF catalyst (Table 1, entries 4 and 5). Also, the RCM reaction of 13 giving a six-membered ring product 14 proceeds in almost a quantitative yield with low loading of tested catalysts Ru7 and Ru9BARF (Table 2, entry 6).
On the basis of these results, we proceeded to extend the scope of the method's applicability to include products with seven-membered rings. In the reaction of compound 15, we noted an 82% yield in the formation of 1-tosyl-2,3,6,7-tetrahydro-1H-azepine (16) with only 1 mol % of Ru9BARF (Table 2, entry 7). In further research, we focused also on pharmaceutically related substrates to emphasize the practical aspect of the emulsion system under study. Therefore, we tested the RCM reaction of allyl (1-methylpent-4-enyl) carbamic acidbenzyl ester (17), which led to precursor of Relacatib-drug with high potency to inhibit cathepsin K that prevents bone resorption (Yamashita et al., 2006). In this case, a very good yield (86%) can be achieved after a 5 h reaction with 1 mol % of Ru9BARF (Table 2, entry 8). Another important example is the synthesis of 20, precursor of a Halidor possessing antispasmodic, vasodilator, and platelet aggregation inhibitor properties (Martin et al., 1974). In this case, reactions also proceed smoothly in neat water using a low loading of catalyst Ru7 ( potentially bind to ruthenium complex and obstruct catalytic cycle. To this aim, diallylated derivatives of ketamine 21 and sildenafil (Viagra) 23 were subjected to the emulsion methodology using Ru9BARF catalyst ( Table 2, entries 10 and 11). It also should be noted that both substrates are solid materials, forming suspensions in water, which could impede their access to the catalyst (also forming an insoluble suspension in water). Therefore, to increase the reagent contact during these RCM reactions, the addition of a small amount of ethyl acetate as a co-solvent was necessary. Despite the presence of an organic co-solvent, the crude products 22 and 24 precipitated from the reaction mixtures and were easily isolated in the high yield by simple filtration and drying.
The positive results of the above experiments encouraged us to investigate the even more difficult crossmetathesis (CM) reactions. The self-cross-metathesis (self-CM) of allylbenzene (25) worked similarly well, as clean formation of the desired product 26 in a high yield was observed using again a relatively low catalyst Ru7 loading of 0.5 mol % (Table 3, entry 1). Similarly, self-CM of functionalized allylbenzenes (phenylpropenoids), eugenol (27), and eugenol acetate (29), resulted in clean formation of the expected products, iScience Article however only in moderate yield (54-56%) despite addition of the co-solvent (Table 3, entries 2 and 3). One of the challenges in the above-mentioned reactions is the known predisposition of the phenylpropenoid substrates to undergo isomerization of the C-C double bond during olefin metathesis (Małecki et al., 2019). It is noteworthy that by-products derived from isomerization processes were not detected under the emulsion conditions, in contrast to reactions carried out in classic organic solvents when only the use of a specialized catalyst prevented by-products formation and gave satisfactory results. As with the RCM reactions, now we also observed that water-soluble substrates such as allyl alcohol (31) were not reactive in the emulsion system when the hydrophobic catalyst was used. In contrast, the CM of lipophilic substrates, such as the reaction between 6-chloro-hexene (33) and a sterically demanding partner (an excess of 4-methyl-1-pentene, 34), proceeded in water in a quantitative yield using 1 mol % of catalysts Ru7 or Ru9BARF (Table 3, entry 5). It has also been found that allylbenzene (25) under acoustic emulsification conditions undergoes a highly efficient CM reaction with cis-1,4-diacetoxy-2-butene (36), without isomerization of 37 to the corresponding styrene derivative (Małecki et al., 2019). Seeing these encouraging results,

OPEN ACCESS
we switched to more challenging CM reactions, including an electron-deficient partners. Under aqueous conditions, this type of CM reaction sometimes leads to lower conversion and selectivity and usually requires higher loading of the olefin metathesis catalyst (Connon and Blechert, 2003). The results obtained for CM reaction of 1-octene (38) with a,b-unsaturated aldehyde 39 showed high efficiency under discussed conditions (Table 3, entry 7).
In a separate study, we have attempted to CM reactions of 9-decen-1-ol (41) with various olefins: 42, 39, and 34. The tested alcohol has an amphiphilic character, which makes it somewhat difficult to carry out the reaction in water. Although the obtained results are in general acceptable, to ensure good selectivity and to avoid the parasitic self-CM process (homodimerization) of alcohol 41, a catalyst loading higher than 2.5 mol % with simultaneous addition of a small volume of the co-solvent was required (Table 4, column A).
To overcome the observed limitation, we decided to further modify the reaction protocol by using the well-established microwave technique (Kirschning et al., 2006;Tagliapietra et al., 2020).

Scope and limitation study of metathesis reactions using microwave irradiation in water
A well-known alternative strategy to perform metathesis reactions of demanding substrates is the use of microwave irradiation (Bantreil et al., 2012;Coquerel and Rodriguez, 2008; Ullah and Arshad, 2017), especially in Table 4. RCM and CM reaction of 9-decen-1-ol (41) with different partners catalyzed by Ru9BARF in the presence of ultrasounds ( )))); column A) a and microwaves (MW; column B) b Entry Substrates Product Catalyst loading (mol %) Yield (%) E:Z ratio c A a )))) B b MW iScience Article expensive and not environmentally benign fluorinated aromatic solvents (Samojłowicz et al., 2011b). In many cases, the use of this thermal activation technique has a general beneficial effect in terms of shorter reaction times, slower catalyst degradation, and less by-products formation (Kappe, 2019;Kappe et al., 2013). Utilization of water as a solvent during experiments in microwave reactor could be particularly profitable, owing to the excellent heat and energy transfer offer by this polar (and green) solvent (Dallinger and Kappe, 2007). Surprisingly, examples describing metathesis reactions in the aqueous media under microwave irradiation conditions are scarce and thus worth of further investigations (Castagnolo et al., 2009;Gułajski et al., 2019). In view of these considerations, we examined the potential synergetic effect of aqueous conditions with microwave irradiation in previously problematic CM reaction of 9-decen-1-ol (41) with olefins 42, 39, and 34. Remarkably, microwave assistance allows for a significant increase of yields and decrease of time in tested reactions ( Table 3, column B). Furthermore, a CM carried out under microwave irradiation does not require the addition of ethyl acetate as cosolvent, which positively affects environmental concerns.
In particular, the CM of alcohol 41 with methyl acrylate (42) using microwave activation allows obtaining the desired product in an 87% yield after only 30 min. In the same reaction performed under standard emulsion conditions, significant amounts of unreacted substrate and homodimerization side-product were observed, with a final yield of 43 that is equal to 57% after 5 h (  (39) and sterically demanding 4-methyl-1-pentane (34). Gratifyingly, for these transformations, even a lower catalyst Ru9BARF loading (1 mol %) was sufficient to achieve a high yield in the microwave reactor (Table 4, entries 2 and 3). The formation of trisubstituted C-C double bonds was accomplished under both ultrasound and microwave conditions ( Table 4, entries [4][5]. Although these reactions gave very good yields under standard emulsion conditions, diene 11 demonstrated a several percent increase in the yield under microwave conditions. Tetrasubstituted C-C double bonds are considered one of the least reactive substrates in olefin metathesis reaction (Heinrich et al., 2020;Lecourt et al., 2018;Mukherjee et al., 2018). Whereas molybdenum alkylidene complexes (Schrock catalysts) are very effective in this case, ruthenium carbenes are less reactive and require specially designed sterically reduced NHC ligands Kuhn et al., 2010;Stewart et al., 2007) or forcing conditions Sashuk et al., 2010;Urbina-Blanco et al., 2013). According to our knowledge, such reactions were never attempted in water. Therefore, we decided to investigate the performance of various Ru-catalysts in RCM, leading to the formation of a tetrasubstituted olefin 47. Unfortunately, quick reconnaissance shown that a specialized ruthenium catalyst with a small NHC ligand (Planer et al., 2020) decomposes very quickly under reaction conditions in water. Similarly, Ru9BARF with a sterically augmented SIPr-like ligand gave no conversions in the case of the formation of a tetrasubstituted C-C double bond. Therefore, we decided to use the SIMes-based catalyst Umicore Grubbs M202 (Ru10), and as a co-solvent, we selected perfluorotoluene (PFT) because it was previously found to give very good results in challenging formation of crowded C-C double bonds by olefin metathesis (Rost et al., 2008;Samojłowicz et al., 2011aSamojłowicz et al., , 2011bSzadkowska et al., 2011). Using microwave irradiation to provide high temperature in the aqueous emulsion system, we conducted a model reaction of diene 46 (Scheme 2) observing 70% of conversion. Interestingly, under precisely the same conditions, but with PFT replaced by ethyl acetate, only 30% conversion was reached, which shows that indeed PFT accelerates this transformation (for details, see Supporting information, Scheme S1). Finally, we were able to run this reaction to completion (96% of 47), however under more forcing conditions (5 3 1 mol % of Ru10, H 2 O/PFT 8:2 v/v, 110 C, 250 W, 50 min). This result shows that water can be used as a diluent also in the formation of tetrasubstituted bonds, further allowing to reduce the volume of PFT typically used (under previously reported conditions, PFT is used as the only solvent in a larger volume (Wang et al., 2021), but here it acts only as a co-solvent added in a small amount of water).
Undoubtedly, these all examples underline the advantages of microwave irradiation in the water-based heterogeneous metathesis system and represent an interesting example of a challenging metathesis reaction using this enabling technique.
Scale-up and application prospects of metathesis in water (use of mechanical stirring and testing practicality of product separation) During the previous screening studies, products were usually isolated by extraction with ethyl acetate and purified by column chromatography. Obviously, such a separation technique is limited only to academic research and small-scale experiments. Interestingly, in many previously discussed reactions in water ll OPEN ACCESS 10 iScience 25, 104131, April 15, 2022 iScience Article emulsions, we observed the formation of oily product droplets on the surface of water, directly after ending the sonication, or after removing the sample from the microwave reactor. This reveals the possibility of avoiding extraction in reactions carried out with water as a diluent on a larger scale. To demonstrate the practical application of the developed method, we attempted the previously tested (cf. Table 3, entry 5) CM reaction of 6-chloro-hex-1-ene (33) with 4-methyl-pente-1-ene (34) in a 150 mL flask using a simple milk frother to form an emulsion (Scheme 3). After the completion of the reaction, the water-insoluble product (35) was isolated by decantation from the post-reaction mixture in 95% purity. Presented examples indeed shows that simple and convenient phase separation works well when water is used as the medium for the metathesis reaction.
When metathesis transformations of pharmaceutical interest substrates are attempted on a larger scale, several practical (technical) aspects need to be addressed, such as optimizing product separation and purification, including rigorous removal of ruthenium traces. In cooperation with a pharmaceutical company, we decided to check if water can be used as a diluent for the olefin metathesis reaction in the pharmaceutical context. To do so, we focused on a preparative synthesis of sildenafil derivative 24 in a larger amount, using techniques and hardware typical for pharmaceutical R + D scale (Scheme 4). In the first step, the Mettler Toledo Easymax workstation reactor equipped with a mechanical stirrer was charged with 17.0 g of a solid sildenafil derivative 23 and water (140 mL). The materials were weighted, and the reactor was charged open to air. This heterogeneous mixture was stirred for several minutes to form a stable suspension of the (water insoluble) substrate, then a solution of (water insoluble) catalyst Ru9BARF in ethyl acetate (35 mL) was added. Next, the reaction mixture was stirred at 40 C for 3 h, until full conversion of the substrate was indicated by TLC analysis. Cooling down of the post-reaction mixture led to the formation of a precipitate that after filtration and drying in the binder vacuum dryer gave 14.5 g of the final product as a beige solid in a 94% yield.
Since ruthenium catalysts have found wide application in the synthesis of active pharmaceutical ingredients (APIs), certain restrictions regarding the heavy metal content in the final product are enforced by national and international regulators. Therefore, the legal limit of transition metal contamination shall typically be less than 10 ppm in the final API (European Medicines Agency, 2008). In addition to ruthenium toxicity, the presence of still-active or even decomposed catalyst causes the risk of isomerization, polymerization, or even degradation of the metathesis product during work-up (Clavier et al., 2007). Whereas several diverse methods for ruthenium scavenging were tested in the context of pharmaceutical production (Vougioukalakis, 2012;Wheeler et al., 2016), according to our knowledge, no attempts to purify a pharmaceutically relevant product obtained in aqueous metathesis have been reported so far. To do so, using the sample of the sildenafil derivative obtained in the previous experiment (Scheme 3), we decided to study this issue. First, using ICP-MS technique, we determined ruthenium contamination in the crude product 10 precipitated from the aqueous emulsion. Then the crude product was divided into four portions and each of them was purified using different protocols; the results are summarized in Table 5. The crude product 24 was found to contain already a reduced level of ruthenium (1,032 ppm) as compared to the initial amount of ruthenium used for reaction ( Table 5, entries 1 and 2). Fortunately, simple recrystallization of the crude product from a mixture of ethyl acetate and ethanol resulted in significant decrease of product contamination ( Table 5, entry 3). An alternative purification protocol, consisting of the use of a ruthenium scavenger developed in our laboratories (Szczepaniak et al., 2019;Toh et al., 2021) (recently commercialized under the trade name SnatchCat), led to 5 ppm of ruthenium residue ( Table 5, entry 4). Interestingly, similar results were obtained after treatment of crude compound 24 (containing 1,032 of ruthenium) with active charcoal (Table 5, entry 5).
The following observations can be noted: (a) interestingly, spontaneous precipitation of product out from the aqueous reaction mixture does not occlude a large amount of catalysts, despite the low solubility in water of the catalyst used; (b) a simple crystallization reduced the ruthenium contamination below the pharmacopoeia accepted levels; (c) a commercially available metal scavenger and active charcoal were similarly effective in further removal of ruthenium level in the crude product. Therefore, it seems that water can be a convenient solvent (diluent) even for solid, polyfunctional substrates exhibiting a high level of molecular complexity, and the ruthenium species formed after catalyst decomposition do not severely contaminate the crude product, which can be purified with relative ease.
It shall be stressed that in both large-scale attempts, we observed almost identical yields compared with the previous small-scale experiments (cf. Table 2, entry 11 and Table 3, entry 5). It should be also noted that despite the high masses of the catalysts used in our studies (Ru7 MW = 1,062 g/mol, Ru9BARF MW = 1,716 g/mol), owing to relatively low loadings, they constitute approximately $4-6 mass-% of the final products. We think that these simple experiments demonstrate the scale-up prospects of the presented methodology, easy product separation and purification, and highlight the utility of water as the reaction medium (or: diluent) in olefin metathesis. We presented a simple synthetic protocol for the preparation of various functionalized products by RCM and CM reactions in a water emulsion system under mild conditions using sonication, microwaves assistance, or mechanical stirring, on air and without using an additional surfactant. In comparison with previously developed protocols for aqueous metathesis reactions, the introduced method applies a relatively low loading of two hydrophobic metathesis catalysts. The studied conditions demonstrate a broad application profile. Moreover, the reactions can be easily scaled up without drop of the yield. In particular, experiments on a larger scale (17 g) have shown the convenience of the elaborated methodology in the synthesis of a pharmaceutically related derivative of sildenafil (Viagra) in an excellent yield and with a low level of ruthenium contamination after simple purification. Therefore, we believe that water as a diluent can find broad applications in olefin metathesis reactions, characterized by a low amount of waste produced, a low loading of the catalyst, and a high selectivity.

Limitations of the study
Despite advantages of the developed methodology, the scope of substrates is limited mainly to the lipophilic non-ionic compounds as for the water-soluble substrates, metathesis reactions were not effective. Moreover, in such cases, the isolation of final product might require either distillation or crystallization.