**1. Introduction**

The employment of renewable feedstocks in the chemical industry is steadily advancing to ensure more efficient use of natural resources, reduce the dependence on fossil raw materials, and give a contribution to achieving sustainable consumption and production patterns [1].

Fats and oils represent an important class of renewable feedstock from which the so-called oleochemicals are obtained. They are abundant in nature, biodegradable, and have nontoxic properties. They have long hydrocarbon chains, resembling the structure of petroleum components, but they are also characterized by several functional groups useful for chemical modification. The major process for transforming fats and oils into oleochemicals is hydrolysis, converting natural triglycerides into crude glycerine and mixtures of fatty acids. The latter are then submitted to reactions involving either the carboxylic group (to afford soaps, esters, amides, amines, and alcohols) or the reduction/oxidation of the C=C double bonds, if present. Among these procedures developed to obtain fine chemicals, the oxidative cleavage of unsaturated fatty acids for the production of dicarboxylic acids, hydroxy acids, and amino acids has received great attention in the last decade [2–4]. Until recently, only two dicarboxylic acids prepared from oleochemicals have been commercialized, i.e., sebacic acid (**1**), obtained by the alkaline cleavage of castor oil [5], and azelaic acid (**2**), which is produced together with

pelargonic acid (**3**) by ozonolysis of oleic acid (**4**) (Figure 1) [6]. Sebacic and azelaic acid are extensively employed in the synthesis of new generation biodegradable copolymers [7]. Azelaic acid, naturally occurring in wheat, rye, and barley, also finds application as an active ingredient in products for the topical treatment of acne [8], and for the stimulation of hair regrowth [9]. It works by inhibiting the growth of skin bacteria causing acne, and by keeping skin pores clear. Pelargonic acid, found in nature as ester derivatives in the oil of pelargonium, is used as an herbicide to prevent the growth of weeds both indoors and outdoors, and as a blossom thinner for apple and pear trees [10].

**Figure 1.** Sebacic acid (**1**), azelaic acid (**2**), pelargonic acid (**3**), and oleic acid (**4**).

Oleic acid is the most abundant monounsaturated fatty acid in nature [11], present in a wide range of vegetable and animal oils and fats. Several works have been published in recent years describing alternative methods to the ozone-promoted oxidative scission, most of which are based on metal catalysis [12]. Among them, some very effective one-pot procedures involve the use of H2O2 as primary oxidant in the presence of tungsten derivatives: (i) methyltrioctylammonium tetrakis(oxodiperoxotungsto)phosphate [13] (40% H2O2, 85 ◦C, yields of compounds **2** and **3** by GC/MS analysis of the crude mixture were 79% and 82%, respectively); (ii) WO3 and Na2SnO3 in *t*-BuOH [14] (31% H2O2, 130 ◦C, sealed glass vial, isolation yields were 89% and 65% for **2** and **3**, respectively); (iii) a new hybrid organic/inorganic polyoxotungstate in *t*-BuOH [15] (30% H2O2, 120 ◦C, yield of compounds **2** and **3** by GC/MS analysis of the crude mixture were 79% and 80%, respectively); iv) H2WO4 [16] (60% H2O2, reflux, isolation yield was 60% for **2**); (v) Na2WO4 aqueous solution/H3PO4 aqueous solution with a suitable phase transfer catalyst in a sealed flask [17] (30% H2O2, 90 ◦C, yield of compound **2** by GC/MS analysis of the crude mixture was max 54%); (vi) the peroxo–tungsten complex [C5H5N(*n*-C16H33)]3{PO4[WO(O2)2]4} as a phase-transfer catalyst/co-oxidant [18] (30% H2O2, 85 ◦C, yields of compounds **2** and **3** by GC/MS analysis of the crude mixture were 79% and 80%, respectively), using in this case oleic acid obtained upon hydrolysis of high oleic sunflower oil by *Candida cylindracea* lipase.

As for biocatalytic methods, only a few reports have appeared in the literature. Song et al. developed [19,20] a multi-step enzymatic procedure (Figure 2) starting from the hydration of oleic acid (**4**), followed by the oxidation of the intermediate alcohol **5** to the ketone derivative **6** that was in turn submitted to Baeyer–Villiger (BV) oxidation to afford ester **7**. The latter was hydrolyzed to afford acid **3** and the hydroxy acid **8**, which was finally oxidized to azelaic acid (**2**).

**Figure 2.** Synthesis of azelaic acid (**2**) from oleic acid (**4**) according to references [19,20].

In [19] the possibility to obtain a regioisomer of ester **7** giving directly diacid **2** upon hydrolysis was described, but it was not considered for further study and optimization in the following publication [20]. The same research group described the preparation of azelaic acid from ricinoleic acid [21]. The use of linoleic acid for the biocatalyzed production of azelaic acid was reported by Hauer et al. [22,23]. A multi-enzymatic one-pot reaction was developed to convert linoleic acid into azelaic acid by combining a 9*S*-lipoxygenase and 9/13-hydroperoxide lyase to obtain 9-oxononanoic acid submitted to the final oxidation to acid **2** catalyzed by an alcohol dehydrogenase. In 2019 the capability of *Candida tropicalis* ATCC20962 to transform nonanoic acid and its esters into azelaic acid **2** with the aid of nonane addition and continuous glucose supply [24] was investigated, to improve the production yield of diacid **2** obtained in the ozonolysis process of oleic acid.

Recently, we were involved in a project aimed at the valorization of the side-stream products generated by an Italian plant for vegetable seed oil refining (Oleificio Zucchi, Cremona) by using biocatalytic methods. An important step of the refining process (neutralization) is represented by the removal of free fatty acids, producing a side-product, called soapstock, which is currently disposed of by Zucchi in bio-digesters. The fatty acids profile of this material depends on the nature of the vegetable oil, and, in particular, the one obtained from sunflower oil is highly enriched in oleic acid [25]. Thus, we started to investigate a novel chemo-enzymatic oxidative scission of oleic acid, to be applied to the sunflower soapstock coming from Oleificio Zucchi for its valorization. The preliminary results of this study were obtained while working on commercial oleic acid, as a model compound, to study each step of the most suitable synthetic procedure more easily using a less complex starting material. The results are herein reported.

#### **2. Results**

The enzymatic synthesis of azelaic acid reported in 2013 by Song et al. [19] (Figure 2) consisted of the use of recombinant *Escherichia coli* cells expressing at the same time the genes encoding an oleate hydratase from *Stenotrophomonas maltophilia*, an alcohol dehydrogenase (ADH) from *Micrococcus luteus*, and a BV monooxygenase (BVMO) from *Pseudomonas putida* KT2440 for the transformation of oleic acid into 9-(nonanoyloxy)nonanoic acid (**7**). The hydrolysis of this latter compound, mediated by a cell extract of *E. coli* expressing the esterase gene from *P. fluorescens*, afforded pelargonic acid (**3**) and 9-hydroxynonanoic acid (**8**). In a further development of the work [20], the oxidation of derivative **8** by an ADH from *P. putida* GPo1 completed the route to azelaic acid. As the final product concentration in the reaction medium was only a few millimolar, likely because of the toxic effects of pelargonic acid on the *E. coli* cells, the authors investigated the hydrolysis of **7** and the subsequent oxidation of derivative **8** into acid **2** by chemical methods [26]. The ester intermediate **7** was purified by extraction and column chromatography, hydrolyzed with sodium hydroxide in methanol/water (4/1) at 60 ◦C to afford 9-hydroxynonanoic acid **8**, which was separated from pelargonic acid by column chromatography. Finally, the oxidation of the terminal hydroxy group of derivative **8** was achieved using NaClO2 (1.2 equiv.), 2,2,6,6-tetramethyl-piperidin-1-yl oxyl (TEMPO) (4 mol%), and NaOCl (2 mol%) in aqueous acetonitrile. After these two steps, no purification was needed. The overall molar yield of azelaic acid from oleic acid was 58%.

We adopted a different strategy (Figure 3), consisting of the epoxidation of oleic acid to derivative **9**, followed by the formation of *threo*-9,10-dihydroxystearic acid (**10**) due to the acid-catalyzed hydrolytic cleavage of the oxirane ring, promoted directly in the epoxidation medium. The chemical oxidation of the diol afforded 9,10-dioxostearic acid (**11**), which was submitted to oxidative cleavage to afford a mixture of pelargonic (**3**) and azelaic (**2**) acid, through the intermediate anhydride **12**. In a recent publication [27], the oxidation of the methyl ester derivative of compound **10**, using a solvent-free procedure of dehydrogenative oxidation catalyzed by commercial 64 wt.% Ni/SiO2 in the presence of 1-decene as a scavenger, was employed to afford a mixture of the two possible regioisomeric vicinal ketols, that were successively cleaved with formic acid/hydrogen peroxide, and afforded up to 80% pelargonic acid and azelaic acid monomethyl ester.

**Figure 3.** The synthesis of azelaic acid (**2**) from oleic acid (**4**) described in this paper. (i) H2O2 35%, Novozyme 435, acetonitrile, 5 h, 50 ◦C; (ii) NaHSO3 saturated solution, H2SO4 2 M, 3 h, r.t.; (iii) atmospheric O2, cat. Fe(NO3)3·9 H2O/TEMPO/NaCl, toluene, 5 h, 100 ◦C; (iv) 35% H2O2, toluene, 3 h, 30 ◦C.

#### *2.1. Epoxidation of Oleic Acid* (**4**) *to 8-(3-Octyloxiran-2-yl)Octanoic Acid* (**9**)

The capability of certain lipases to catalyze the perhydrolysis (i.e., lysis by hydrogen peroxide) of carboxylic acid esters, hence forming peroxycarboxylic acids in aqueous hydrogen peroxide solutions, had been already patented by Clorox co. in the late eighties [28]. In 1990, an immobilized form of lipase B from *Candida antarctica* (Novozyme 435) was shown [29] to catalyze the formation of peroxycarboxylic acids directly from the corresponding carboxylic acid. In that case, the reaction was combined with the epoxidation of alkenes. A few years later, Warwel et al. [30] described that, when unsaturated fatty acids (or their esters) are treated with hydrogen peroxide in the presence of Novozyme 435, epoxidized derivatives are obtained through two sequential steps. Firstly, unsaturated fatty acids are converted into unsaturated peroxy acids by lipase-catalyzed perhydrolysis. Unsaturated peroxy or carboxylic acids are in turn epoxidized via a classical Prileshajev reaction, which is, in this case, referred to as a "self-epoxidation reaction" even though it proceeds predominantly via an intermolecular process.

The reaction has been widely exploited not only for the epoxidation of fatty acids and esters but also for the derivatization of vegetable oils [31]. Typically, the reaction medium consists of an aqueous layer containing hydrogen peroxide, an organic layer (usually a toluene solution) containing the fatty acid derivative, and a solid phase represented by the immobilized enzyme. The main issue of this chemo-enzymatic procedure is the deactivation of lipase. Temperature, reaction time, concentration of H2O2 in the reaction medium, and the related concentration of peracid generated in situ are critical parameters to be considered. Temperatures not higher than 50 ◦C, diluted H2O2 solution (max 1% *w*/*w* in the final solution) and reaction times not longer than 6 h represent the most common experimental conditions.

We decided to perform the chemo-enzymatic epoxidation of oleic acid in a water-miscible solvent, such as acetonitrile, for the following reasons: i) to promote the dissolution of both oleic acid and H2O2 in the same medium, and ii) to enable the in situ acid-catalyzed hydrolysis of the epoxide derivative at the end of the reaction, after removal of the enzyme, by addition of a diluted solution of sulfuric acid. The preliminary experiments were carried out with 0.30 mmol of commercial oleic acid (91% purity by GC/MS, the major contaminants are palmitic and stearic acid), changing the molar ratio H2O2/oleic acid (1.8 and 2.2), the temperature (30 ◦C and 50 ◦C), the amount of Novozyme 435 (10 mg and 30 mg), and the solvent volume (2 mL and 6 mL). The reactions were monitored by GC/MS. The results of this screening are reported in Tables S1 and S2 (see Supplementary Material). The following conditions were found to be optimal for running the reaction: 0.15 M oleic acid, 0.27 M H2O2, 5 mg·mL−<sup>1</sup> Novozyme 435, in acetonitrile at 50 ◦C for 5 h under stirring with final 98% conversion (GC/MS,). Epoxide **9**

could be recovered from the reaction mixture at 83% isolation yield (see Supplementary Material), starting from commercial oleic acid. The stereochemistry of the oxirane ring of derivative **9** was based on the *cis* configuration of oleic acid and was confirmed by comparison with literature data (see Supplementary Material).
