*2.3. Oxidation of 9,10-Dihydroxystearic Acid* (**10**) *to 9,10-Dioxostearic Acid* (**11**)

The following procedures were tested to achieve the oxidation of either one or both the hydroxy groups of derivative **10**: (i) alcohol dehydrogenase—mediated oxidation (commercial kit from EVOXX); (ii) chemo-enzymatic oxidation with laccase and hydroxybenzotriazol (HOBt); (iii) aerobic oxidation with catalytic Fe(NO3)3·9 H2O, TEMPO, and NaCl. Only the latter was successful and diol **10** could be converted into the corresponding dioxo derivative **11** (Figure 3). Starting from 0.50 g of diol **10**, according to the literature [32], a loading of 1 mol% for each catalyst was enough to afford complete conversion into the diketone in toluene solution at 100 ◦C in 5 h. After work-up (quenching with water and extraction), the crude residue was submitted directly to the following step of oxidative cleavage.

#### *2.4. Oxidative Cleavage of 9,10-Dioxostearic Acid* (**11**) *to Azelaic* (**2**) *and Pelargonic Acid* (**3**)

For the final step of the synthetic procedure, we investigated the Baeyer–Villiger (BV) oxidation of diketone **11**, to prepare the corresponding anhydride **12**, and hydrolyze it to azelaic (**2**) and pelargonic acid (**3**). A wide range of oxidants has been employed for the BV reaction, including mineral and organic peracids. Hydrogen peroxide can be used if suitably activated by a catalyst, or in the presence of a strong acid, or even in alkaline conditions [33,34]. α-Diketones react readily with BV reagents: in inert solvents anhydrides are formed, while in alkaline or acidic media simple carboxylic acids are generally produced in good yields [35]. In 1930 Böeseken et al. [36] prepared 9,10-diketostearic acid (**11**) by oxidation of 9-octadecynoic acid with 70% nitric acid at 10–25% yield and submitted it to the reaction with 15% excess peracetic acid in acetic acid for one day. They obtained the quantitative conversion of the diketone into a mixture of acids **2** and **3.**

Thus, we first considered the possibility to perform the BV oxidation of compound **11** with the corresponding peroxycarboxylic acid produced by lipase-mediated perhydrolysis in the presence of H2O2, using for preliminary experiments a sample of **11** isolated and purified by column chromatography. We treated dioxostearic acid **11** (50 mg) with 1.6 mol of H2O2 per mol of dioxostearic acid in the presence of 2.5 mg·mL−<sup>1</sup> Novozyme 435 in toluene (2 mL), the same solvent employed for the preceding oxidation. The cleavage was complete after 3 h at 30 ◦C, to give a mixture of 96% pelargonic and azelaic acids, with tiny quantities of octanoic (0.3%), stearic (0.4%), and palmitic acid (2%), with other minor components (GC/MS). To our surprise, when the blank reaction was carried out in parallel in identical conditions without the presence of the enzyme, the same oxidative cleavage was observed, affording a mixture containing acids **2** and **3**, besides 51% of intermediate anhydride **12** ( 1H NMR). The presence of intermediate **12** was highlighted by NMR spectroscopy. The 13C NMR

spectrum of the crude reaction mixture showed the presence of two singlets at 169.73 and 169.67 ppm for the carboxylic carbon atoms of the anhydride, next to those around 180 ppm which belong to acids **2** and **3** and to the COOH group of compound **12.** In the 1H NMR spectrum, the triplet of the CH2 groups linked to the CO-O-CO moiety is at 2.44 ppm, a little more deshielded than the triplet of the CH2 beside COOH in compounds **2** and **3** and in the anhydride itself, occurring at 2.35 ppm. The 1H and 13C NMR spectra of anhydride **12** are not known in the literature, and the spectroscopic data of lauric anhydride reported in [37] were used as reference data.

The reaction was repeated in acetonitrile and only 11% of anhydride was found in the final mixture. When the oxidation was carried out in toluene and2MH2SO4 was added to the reaction mixture during the workup procedure, after having decomposed peroxy species with NaHSO3 saturated solution, complete hydrolysis of intermediate **12** was obtained.

After investigation of every single step, the whole procedure was performed starting from 2 g of commercial oleic acid. Acidic hydrolysis was performed soon after epoxidation in a one-pot procedure, to afford diol derivative **10** as a pure compound at 70% isolated yield by filtration of the first crop of crystalline material and recovery of other product by further treatment of the mother liquors. The oxidation to dioxoderivative **11** gave a crude compound (75% purity by GC/MS) that was submitted directly to the last step of oxidative cleavage in toluene with only 35% H2O2, to provide a mixture of azelaic and pelargonic acids. Diacid **2** was recovered following a procedure, which had been already described in the literature [16], and based on the solubility of compound **2** in hot water. Repartition between ethyl acetate and hot water afforded an aqueous phase from which azelaic acid crystallised upon cooling. After three extraction cycles, diacid **2** could be recovered as a pure compound in 73% yield. Pelargonic acid **3** was isolated from the organic phase at 77% isolation yields, showing 91% chemical purity (GC/MS). The separation of diacid **2** from compound **3** was also investigated by using column chromatography, eluting with hexane–EtOAc mixtures with an increasing amount of the more polar solvent (see Supplementary Material), affording pure **2** and **3** in slightly higher isolation yields (81% and 84%, respectively).

## **3. Discussion**

Ozonolysis of alkene bonds is a useful chemical transformation which is employed not only at the laboratory level but also at industrial scale for the rapid and effective oxidative cleavage of C=C double bonds [38]. The primary concern related to ozonolysis chemistry is represented by the serious safety issues connected with the reaction, and in particular with the explosive hazard due to the instability of intermediate ozonides. The present industrial production of azelaic acid is entirely based on ozonolysis of oleic acid, being the global azelaic acid market valued at 94 million USD in 2017 and expected to reach 140 million USD by 2025 [39]. The market is mainly driven by growing demand for plastics and lubricants, which hold above 70% of global azelaic acid consumption.

The growing attention towards the development of safer and more environmentally friendly production technologies has stimulated the investigation of alternative methods for the conversion of oleic acid into azelaic acid. Suitable references have been reported in the Introduction. We gave a contribution to this search by investigating a chemo-enzymatic approach to achieve the target oxidative scission.

We decided to use the in-situ peroxidation of oleic acid **4** by lipase-mediated perhydrolysis in the presence of hydrogen peroxide 35% as a safe procedure to afford the peroxycarboxylic acid needed to promote the epoxidation step at the beginning of the synthetic sequence. Oleic acid itself undergoes the conversion into the reactive peroxy species, so it is possible to avoid the use of an additional carboxylic acid that would remain in the reaction mixture as a by-product to be removed from the desired final compound. The advantage of the proposed procedure is also that storage and manipulation of peracid are avoided: it is generated in the reaction medium, and the excess is destroyed at the end of its use. The best biocatalyst for this reaction is Novozyme 435 which has the advantage of being an immobilized form of *C. Antarctica* B that can be recovered and re-used. Preliminary experiments were performed

starting from 1 g of oleic acid, and recovering the enzyme by filtration and washing with water and acetonitrile. The enzyme was kept at 4 ◦C for 18 h and re-used in a subsequent reaction. After four runs the conversion of oleic acid (GC/MS analysis) into the epoxide was 78%. The evaluation of the enzyme reusability in these experiments can only be used as a first orientation because it is influenced by the effects of manipulation and storage on the enzyme support, overlapping those due to hydrogen peroxide and peracid. Long-term performance of the enzyme should be best studied under continuous process conditions, as suggested by recent literature [40]. Very positive results on the stability of this lipase in this type of reaction have been obtained using both a packed-bed reactor [41] and a continuous stirred tank reactor [40]. This kind of investigation is now in progress in our research group.

We considered also the possibility to avoid the isolation and purification of some of the intermediates of our procedure to reduce quantities of waste, solvents and separation aids. We chose the solvent of the epoxidation reaction to telescope the first two steps of the procedure, and perform the acid-catalyzed opening of the oxirane ring without isolation of epoxide **9**. Diol derivative **10** could be obtained as a pure crystalline compound by crystallization from acetonitrile, after quenching the peroxy species with NaHSO3 and promoting epoxide cleavage with catalytic H2SO4 aqueous solution. No column chromatography was needed, thus favoring the isolation yield of diol **10** and limiting further use of solvents. Even the purification of dioxostearic acid **11** could be avoided, and the raw material was submitted directly to the following step to afford azelaic and pelargonic acid. The use of toluene as a solvent for both the diol oxidation and the final oxidative cleavage reaction will be useful in the future for developing the procedure in continuous flow mode.

For the last step of the whole sequence, we discovered the unexpected capability of H2O2 35% w/w to promote the oxidative cleavage of diketone **11** in organic solvents, either toluene or acetonitrile, at 30 ◦C, without the addition of any catalyst. H2O2 is considered as a green oxidant, generating water as a by-product. It is safely stored and transported, and easily available on the market at a cheap price.
