Avoiding hot-spots in Microwave-assisted Pd/C catalysed reactions by using the biomass derived solvent γ-Valerolactone

Herein, we report the use of γ-valerolactone as a new biomass-derived reaction medium for microwave assisted organic synthesis. The interaction of this solvent with microwaves and its heating profile under microwave irradiation has been fully characterized for the first time, demonstrating its stability and the applicability in microwave assisted Pd/C catalysed reactions avoiding the arcing phenomena frequently observed in these conditions. The use of γ-valerolactone demonstrated to be compatible with aliphatic and aromatic amines in the hydrogen transfer Pd/C mediated synthesis of benzimidazoles.

The technology used for promoting a synthetic transformation plays a crucial role in the definition of its chemical efficiency and sustainability. In particular, effective heating represents a key challenge for accessing the desired control on the chemical reactivity while ensuring an adequate energy consumption. While flow reactors play certainly a key role for ensuring exceptionally fast heat and mass transfer compared to batch reactors 1 , microwave (MW) dielectric heating still represents nowadays one of the most efficient mean for heating up a reaction mixture and potentially reduce energy consumption 2 and combine the need for effective synthetic procedures especially on a larger scale [3][4][5][6] . The discussion about how MWs are able to accelerate a reaction is still open [7][8][9] . Even though in most cases the postulated microwaves effects demonstrated to be just related to the erroneous temperature measures [10][11][12][13][14][15] or to a misinterpretation of the data obtained 16 , in the last few years the evidence of catalytic MW effects has been observed and demonstrated by different groups especially in reactions involving heterogenous catalysts 8,9,17,18 . Many problems related to reproducibility, predictability and safety issues using MWs have been overcame during the last 30 years, thus transforming MWs in a mature, useful and ordinary technology routinely applied in most organic and medicinal chemistry laboratories as well as in material synthesis and in industrial production [4][5][6] . Nevertheless, in MW assisted reactions hot spot formation remains a main safety and efficiency problem, responsible for explosions and loss of materials, still remain unsolved right now. Hot spot formation occur with arching phenomena during MW irradiation. They have been widely studied and represent a well documented phenomenon specially occurring because of differential heating of solid catalysts under MW irradiation 9,[19][20][21][22] . Even though explosions do not occur during the MW irradiation, arching phenomena are responsible for loss of efficiency of the catalytic process especially when charcoal supported metal are used as the catalysts 9,[19][20][21][22] . The problem has been efficiently overcome in industrial processes by designing dedicated MW reactors 2,19,21 operating at different frequencies (e.g. 5.8 GHz) than those commonly used in commercially available MW apparata for organic synthesis (2.45 GHz). Alternatively, another solution proposed is to expose the reaction only to magnetic field instead of the electromagnetic one. Nevertheless, it is still not trivial to avoid hot spot formation in traditional MW reactors especially when Pd/C is used as the catalyst in solvents not able to adbsorbe the electromagnetic irradiation (e.g. toluene). Thus hot spot represent one of the main limitations to the use of MW dielectric heating by some organic chemists. It is well documented 20,22,23 and we directly experienced 24-26 that hots spots are usually observed when solvents with low boiling points are used in the presence of a heterogeneous catalysts under high electric fields. Sometimes, this issue can be circumvented in lab scale microwave assisted reactions by the use of solvents with higher boiling points than the temperature requested by

Results
As GVL properties under MW irradiation are completely unexplored, we firstly recorded the heating profiles under MW dielectric heating of GVL to evaluate its adequacy for MW assisted reactions. At this aim, 4 mL of GVL were irradiated for 10 min at 50, 100, 150, and 200 Watt fixed power respectively, and the heating profiles recorded have been summarized in Fig. 1a. It is interesting to highlight that the curves obtained are almost identical for all the four conditions tested while the temperature measured are proportional to the power used (see Table S2). The same MW irradiation conditions were applied to water, DMF, toluene, and NMP for a better characterization of the GVL heating profiles with respect to already characterized solvents (Fig. 1b). GVL interacts with MW in a comparable manner than NMP, the only main difference being represented by the higher final temperature reached with GVL. The biomass-based solvent, demonstrated to be a strongly MWs absorbing compound able to reach very high temperatures in short times, and without the need for the common long ramp times requested for example by toluene or water. It is interesting to note that no degradation of GVL was observed even after irradiation for longer than 10 min at 200 W, while NMP decomposes after 10 min at 150 and/or 200 W.
These data suggest that GVL is a valuable candidate for becoming a really good medium to define safe and efficient MW assisted processes. To demonstrate this thesis, we focused on the benzimidazole synthesis protocol as a representative hydrogen transfer reaction. Infact, this process proved to be very sensitive to the reaction medium and only toluene allowed the proper chemical efficiency to prepare the desired products but leaving room for continuous dangerous formation of hot spot and consequent explosions. This process appers to be ideal for investigating the efficiency in the use of GVL under MWs irradiation including its influence in the hot spot formation.
Nevertheless, as GVL is a lactone, its chemical stability in the presence of an amine [68][69][70] , which is one the reactant participating the benzimidazole synthesis makes this process even more challenging for GVL and for this reason in our opinion, even more worthy to be investigated.
We satisfactorily achieved a 90% conversion to 3a (measured by GC analysis) without detecting any traces of the intermediate 4a or of GVL degradation by-products. With this promising data in hand, we have also found that by using 0.35 equivalents of Et 3 N there was no impact on the conversion (Fig. 2, entry 2), proving that two of the three groups of the tertiary amine could be transferred to 1a 25 25 lower conversions were observed. Irradiation of 1a in GVL led to good reaction yields also when the catalyst amount was reduced to 5 or even 2.5 mol% (Fig. 2, entries 8-9), while when 1 mol% was used conversion reached only 70% (Fig. 2, entry 10). The same protocol has been extended to Bu 3 N obtaining also in this case very good conversions and suggesting a more general applicability to tertiary amines ( Fig. 2, entry 11). In none of the reactions using GVL, the formation of hot spots was detected, while when toluene was used as medium ( Fig. 2, entry 7), the process needed to be repeated 3 times to obtain the reported results and in two out of the three experiments arching phenomena destroyed the reaction tube. Starting from these promising results, we were particularly intrigued about the possibility to study the methodology using primary amines in the presence of GVL. Therefore, 1a was irradiated in the presence of 1 equivalent of BuNH 2 (2a) in the same reaction conditions reported in the optimized conditions (Fig. 2, entry 9: 2.5 mol % Pd/C, crotonitrile (2 eq.), AcOH (0.1 eq.), MW, 170 °C, 20 minutes). Under these conditions, benzimidazole 3b was formed in 65% yield accompanied by the presence of 16% of the by-product 5a resulting from the nuclephilic attack of BuNH 2 on GVL (Fig. 2). Although this result was expected [68][69][70] we further investigated this transformation using higher quantity of Pd/C (5 mol%) at different temperatures always observing an increasing amount of by-product 5a (Fig. 3). Promising results have been finally obtained using dry GVL (molecular sieves) with the formation of only 15% of 5a and a 75% conversion into 3b. The best results have been achieved by irradiating in dry GVL, with 5 mol% of Pd/C, 2 equivs of crotonitrile at 170 °C for 20 minutes in the absence of AcOH. In the previously developed procedure in toluene, AcOH demonstrated to be crucial for an efficient synthesis of benzimidazoles via the hydrogen transfer protocol, while in the case of GVL this acid just catalysed the lactone ring opening lowering the conversion to desired product 3b. Different o-phenylendiamine derivatives and primary amines have been treated in the same reaction conditions and it is interesting to note that good results in term of both conversions and isolated yields are observed using phenethylamine 2b as well as p-methoxyphenethylamine 2c (Fig. 4, entry 1-2) 25,71 . The same reaction product (3e) was obtained using both amines 2d (Fig. 4, entry 3) and 2e (Fig. 4, entry 4) thus demonstrating the incompatibility of this protocol with double bonds even using excess of crotonitrile as the hydrogen acceptor 72 . This finding was confirmed by the experiments with allyl amine that furnished the 3 f in 68% yield (Fig. 4, entry 5).
The reaction yields dramatically lowered down using hexadecan-1-amine (Fig. 4, entry 6) probably because of some solubility problems observed during the transformation, while a great 80% yield is observed with piperonylamine 2 h (Fig. 4, entry 7). Very good results were obtained using 2-picolylamine obtaining with an almost quantitative conversion, the highly valuable benzimidazole derivative 3i (Fig. 4, entry 8) used for its luminescent properties in dyes-synthesizer solar cells 73 and as a metals scavenger because of its ability to complex them. The amine used seems to do not impact in the outcome of the reaction. Otherwise the phenylendiamine derivative used plays a key role in this transformation that furnished lousy results in the presence of halogens (Fig. 4, entry 9).
A Pd mediated dehalogenation of the aromatic ring contemporary with GVL ring opening by the amine used was always observed, indicating that the starting halogenated phenylendiamine derivatives are not nucleophilic enough to react with the intermediate imine formed. Excellent results were observed   (Fig. 4, entry 10-11), while (3,4-diaminophenyl)(phenyl)methanone (1d) and 4-methylbenzene-1,2-diamine (1e) furnish only acceptable yields especially with butylamine (Fig. 4, entry 12-15) 74 . The protocol developed consents to obtain benzimidazole derivatives in a more efficient way both in term of reaction yields (3c, 3d, 3h, 3l), times (20 min versus 90 min), and safety issues (no hot spots observed using GVL) with respect to the original protocol 25 .
The use of GVL for avoiding arching phenomena in Pd/C catalyzed MW assisted transformations was also investigated using Heck 44 and Sonogashira 45 cross couplings, and in reduction/hydrogenolysis as model 75 (Fig. 5). The expected products were always obtained in very good yields without any explosion or hot spot formation observed, thus demonstrating the general applicability of GVL in microwave assisted Pd/C catalysed reactions.

1,2-diphenylethyne (8).
To a solution of iodobenzene (57 µL, 0.5 mmol), and ethynylbenzene (84 µL, 0.75 mmol) DABCO (70 mg, 0.6 mmol) in GVL (1 mL), Pd/C 10 wt% (5.3 mg, 0.0025 mmol) was added. The mixture was irradiated with MWs for 10 min at 60 °C (max internal pressure 200 psi). Petroleum ether was added and the reaction mixture was filtered over celite pad washing with water and 1 M HCl. The organic layer was dried over dry Na 2 SO 4 and the solvent was removed under reduced pressure. The crude oil was purified by flash chromatography (petroleum ether) to obtain 8 as a white solid (71 mg (S)-2-amino-8-oxodecanoic acid (10). To a solution of 9 (100 mg, 0.16 mmol) in dry GVL (1 mL) Pd(OH) 2 /C 10%wt (22 mg, 0.016 mol) was added. The 10 mL vial was introduced into the Discover Microwave Synthesizer and purged three times with vacuum/H 2 and finally charged with H 2 (6.8 atm). The mixture was irradiated with MWs microwave irradiation for 30 min at 100 °C (max internal pressure 200 psi). The vial was vented, flushed with nitrogen, and the reaction mixture was filtered over celite pad washing with MeOH. The solvent was evaporated and the cure mixture purified by flash chromatography using AcOEt:MeOH (9:1) obtaining 49 mg of 10 (75% yield). 1