Advances for Biorefineries: Glycerol Hydrogenolysis to 1,3-Propylene Glycol

Humanity’s growing dependence on non-renewable resources and the ensuing environmental impact thus generated have spurred the search for alternatives to replace chemicals and energy obtained from petroleum derivatives. Within the group of biofuels, biodiesel has managed to expand worldwide at considerable levels, going from 20 million tn/year in 2010 to 47 million tn/year in 2022, boosting the supply of glycerol, a by-product of its synthesis that can be easily used as a renewable, clean, low-cost raw material for the manufacture of products for the chemical industry. The hydrogenolysis of glycerol leads to the production of glycols, 1,2-propylene glycol (1,2-PG) and 1,3-propylene glycol (1,3-PG). In particular, 1,3-PG has the highest added value and has multiple uses including its application as an additive in the polymer industry, the manufacture of cosmetics, cleaning products, cooling liquids, etc. This review focuses on the study of the hydrogenolysis of glycerol for the production of 1,3-PG, presenting the main reaction mechanisms and the catalysts employed, both in liquid and vapor phase. Engineering aspects and the effect of the operating variables to achieve maximum yields are discussed. Finally, studies related to the stability and the main deactivation mechanisms of catalytic systems are presented.


Introduction
For some years now, society has become aware of the environmental impact generated by the use of fossil resources to obtain energy sources and products for the chemical industry. Thus, the search for alternative sources has been encouraged to replace, in part, the current dependence on non-renewable sources.
In this context, the concept of biomass arises, defined by the technical specification of the European Committee for Standardization (CEN), as any material of biological origin except those belonging to the field of geology that have undergone a mineralization process. Biomass transformation processes, unlike fossil fuels, do not generate toxic residues or pollutants to the environment, and the CO 2 emissions derived from their combustion result in a neutral carbon balance.
Chemical products derived from biomass come from plant biopolymers such as cellulose, hemicellulose, lignin, and starch. From these products, it is possible to obtain different groups of monosaccharides (hexoses and pentoses), carboxylic acids, phenols, and alcohols, among others, depending on the nature of the polymer [1].
With respect to energy sources, they are based on biofuels that have been produced from plant crops (first-generation biomass), plant or animal waste derived from human activities (second-generation biomass), or by fermentative biological processes (thirdgeneration biomass) [2]. For example, bioethanol and biodiesel have experienced increased production in the last decade [3].
In particular, for more than ten years now, biodiesel has undergone the greatest growth worldwide. The expansion of its production has been observed not only in developed In particular, for more than ten years now, biodiesel has undergone the greatest growth worldwide. The expansion of its production has been observed not only in developed countries such as Germany, Italy, France, and the United States, but also in developing countries such as Brazil, Argentina, Indonesia, and Malaysia [4]. A recent report by the Food and Agriculture Organization of the United Nations (FAO) forecasts that by 2026, production could reach 40 million tons per year, 12% more than in 2016 [5].
In Argentina, biodiesel has been added as another link in the value-added chain of the national agro-industrial complex since 2007, placing Argentina as the fifth large-scale technology producer in the world and among one the most efficient in the world. [6].
From a chemical point of view, biodiesel consists of a mixture of esters from longchain fatty acids, produced from a catalytic transesterification reaction, using virgin or used vegetable oils or animal fats and short-chain alcohols such as methanol (Scheme 1), ethanol (Scheme 2), or 1-propanol (Scheme 3). Alkali homogenous catalysts are conventionally used for biodiesel production such as NaOH, KOH, NaOCH3, and KOCH3, but also heterogeneous catalysts such as acid solid catalysts (e.g., zeolites, clays, mixed metal oxides, sulfated oxides, sulfonated carbon-based materials, and cation exchange resins) and alkali solid catalysts (e.g., alkali-earth metal oxides, hydrotalcites, alkali zeolites, and supported alkali metals) [7]. For every 10 kg of biodiesel produced, 10 kg of triglyceride source and 1 kg of alcohol are required, generating 1 kg of crude glycerol as a by-product with a purity that varies in the range of 55-90% due to the presence of triglyceride, alcohol, soaps, and catalyst residues, among others [8].  Glycerol is a polyol with three hydroxyl groups and interesting physicochemical properties that allow its application in different industries. Its main application is in the pharmaceutical industry for the production of medicinal drugs (~18%), followed by its use in the cosmetics industry (~16%), polymers (~14%), food (~11%), and in the chemical industry for the manufacture of triacetin (~10%), alkyd resins (~8%), tobacco (~6%), detergents (~2%), cellophane (~2%), and explosives (~2%) (Figure 1) [9].   REVIEW 2 In particular, for more than ten years now, biodiesel has undergone the greatest growth worldwide. The expansion of its production has been observed not only in developed countries such as Germany, Italy, France, and the United States, but also in developing countries such as Brazil, Argentina, Indonesia, and Malaysia [4]. A recent report by the Food and Agriculture Organization of the United Nations (FAO) forecasts that by 2026, production could reach 40 million tons per year, 12% more than in 2016 [5].
In Argentina, biodiesel has been added as another link in the value-added chain of the national agro-industrial complex since 2007, placing Argentina as the fifth large-scale technology producer in the world and among one the most efficient in the world. [6].
From a chemical point of view, biodiesel consists of a mixture of esters from longchain fatty acids, produced from a catalytic transesterification reaction, using virgin or used vegetable oils or animal fats and short-chain alcohols such as methanol (Scheme 1), ethanol (Scheme 2), or 1-propanol (Scheme 3). Alkali homogenous catalysts are conventionally used for biodiesel production such as NaOH, KOH, NaOCH3, and KOCH3, but also heterogeneous catalysts such as acid solid catalysts (e.g., zeolites, clays, mixed metal oxides, sulfated oxides, sulfonated carbon-based materials, and cation exchange resins) and alkali solid catalysts (e.g., alkali-earth metal oxides, hydrotalcites, alkali zeolites, and supported alkali metals) [7]. For every 10 kg of biodiesel produced, 10 kg of triglyceride source and 1 kg of alcohol are required, generating 1 kg of crude glycerol as a by-product with a purity that varies in the range of 55-90% due to the presence of triglyceride, alcohol, soaps, and catalyst residues, among others [8].  Glycerol is a polyol with three hydroxyl groups and interesting physicochemical properties that allow its application in different industries. Its main application is in the pharmaceutical industry for the production of medicinal drugs (~18%), followed by its use in the cosmetics industry (~16%), polymers (~14%), food (~11%), and in the chemical industry for the manufacture of triacetin (~10%), alkyd resins (~8%), tobacco (~6%), detergents (~2%), cellophane (~2%), and explosives (~2%) (Figure 1) [9].   REVIEW 2 In particular, for more than ten years now, biodiesel has undergone the greatest growth worldwide. The expansion of its production has been observed not only in developed countries such as Germany, Italy, France, and the United States, but also in developing countries such as Brazil, Argentina, Indonesia, and Malaysia [4]. A recent report by the Food and Agriculture Organization of the United Nations (FAO) forecasts that by 2026, production could reach 40 million tons per year, 12% more than in 2016 [5].
In Argentina, biodiesel has been added as another link in the value-added chain of the national agro-industrial complex since 2007, placing Argentina as the fifth large-scale technology producer in the world and among one the most efficient in the world. [6].
From a chemical point of view, biodiesel consists of a mixture of esters from longchain fatty acids, produced from a catalytic transesterification reaction, using virgin or used vegetable oils or animal fats and short-chain alcohols such as methanol (Scheme 1), ethanol (Scheme 2), or 1-propanol (Scheme 3). Alkali homogenous catalysts are conventionally used for biodiesel production such as NaOH, KOH, NaOCH3, and KOCH3, but also heterogeneous catalysts such as acid solid catalysts (e.g., zeolites, clays, mixed metal oxides, sulfated oxides, sulfonated carbon-based materials, and cation exchange resins) and alkali solid catalysts (e.g., alkali-earth metal oxides, hydrotalcites, alkali zeolites, and supported alkali metals) [7]. For every 10 kg of biodiesel produced, 10 kg of triglyceride source and 1 kg of alcohol are required, generating 1 kg of crude glycerol as a by-product with a purity that varies in the range of 55-90% due to the presence of triglyceride, alcohol, soaps, and catalyst residues, among others [8].  Glycerol is a polyol with three hydroxyl groups and interesting physicochemical properties that allow its application in different industries. Its main application is in the pharmaceutical industry for the production of medicinal drugs (~18%), followed by its use in the cosmetics industry (~16%), polymers (~14%), food (~11%), and in the chemical industry for the manufacture of triacetin (~10%), alkyd resins (~8%), tobacco (~6%), detergents (~2%), cellophane (~2%), and explosives (~2%) (Figure 1) [9].   However, the crude glycerol obtained in biodiesel synthesis has a low level of purity and is not suitable for direct application in the above markets, thus requiring treatment and purification [10]. A fractional distillation process allows technical grade glycerol to be obtained but neither the pharmaceutical nor agri-food market can absorb the entire supply of glycerol. The surplus has oversaturated these markets, devaluing the price of crude glycerol. In the last 10 years, the price of crude glycerol has been between 200 and 400 USD/ton, with the current world average price being in the order of 480 USD/ton ( Figure 2) [11]. Due to its wide availability and low cost, the search for new alternatives for the transformation of glycerol is very attractive [12]. The chemical functionality of the glycerol molecule makes it a platform molecule for obtaining chemical products and biofuels ( Figure  3). Other chemical products can be obtained from glycerol through oxidation [6], nitration [9], polymerization [9], esterification [13], etherification [14], reforming [15], chlorination [16], carbonation [17], dehydration [18], and hydrogenolysis [19]. However, the crude glycerol obtained in biodiesel synthesis has a low level of purity and is not suitable for direct application in the above markets, thus requiring treatment and purification [10]. A fractional distillation process allows technical grade glycerol to be obtained but neither the pharmaceutical nor agri-food market can absorb the entire supply of glycerol. The surplus has oversaturated these markets, devaluing the price of crude glycerol. In the last 10 years, the price of crude glycerol has been between 200 and 400 USD/ton, with the current world average price being in the order of 480 USD/ton ( Figure 2) [11]. However, the crude glycerol obtained in biodiesel synthesis has a low level of purity and is not suitable for direct application in the above markets, thus requiring treatment and purification [10]. A fractional distillation process allows technical grade glycerol to be obtained but neither the pharmaceutical nor agri-food market can absorb the entire supply of glycerol. The surplus has oversaturated these markets, devaluing the price of crude glycerol. In the last 10 years, the price of crude glycerol has been between 200 and 400 USD/ton, with the current world average price being in the order of 480 USD/ton ( Figure 2) [11]. Due to its wide availability and low cost, the search for new alternatives for the transformation of glycerol is very attractive [12]. The chemical functionality of the glycerol molecule makes it a platform molecule for obtaining chemical products and biofuels ( Figure  3). Other chemical products can be obtained from glycerol through oxidation [6], nitration [9], polymerization [9], esterification [13], etherification [14], reforming [15], chlorination [16], carbonation [17], dehydration [18], and hydrogenolysis [19]. Due to its wide availability and low cost, the search for new alternatives for the transformation of glycerol is very attractive [12]. The chemical functionality of the glycerol molecule makes it a platform molecule for obtaining chemical products and biofuels ( Figure 3). Other chemical products can be obtained from glycerol through oxidation [6], nitration [9], polymerization [9], esterification [13], etherification [14], reforming [15], chlorination [16], carbonation [17], dehydration [18], and hydrogenolysis [19].   Of all of the processes mentioned, catalytic hydrogenolysis is an interesting alternative due to the high oxygen content in the glycerol molecule (O/C = 1). In this sense, the breaking of C-O bonds of glycerol leads to the formation of glycols such as 1,2-propylene glycol (1,2-PG) and 1,3-propylene glycol (1,, and C3 alcohols such as 1-propanol (1-POH) and 2-propanol (2-POH), although the breaking of C-C bonds is also present, which allows for the generation of other low molecular weight alcohols such as ethanol (EtOH) and methanol (MeOH). Of these, glycols are the most valuable products, as shown in Figure 4 [20].
Reactions 2022, 3, FOR PEER REVIEW 5 Of all of the processes mentioned, catalytic hydrogenolysis is an interesting alternative due to the high oxygen content in the glycerol molecule (O/C = 1). In this sense, the breaking of C-O bonds of glycerol leads to the formation of glycols such as 1,2-propylene glycol (1,2-PG) and 1,3-propylene glycol (1,, and C3 alcohols such as 1-propanol (1-POH) and 2propanol (2-POH), although the breaking of C-C bonds is also present, which allows for the generation of other low molecular weight alcohols such as ethanol (EtOH) and methanol (MeOH). Of these, glycols are the most valuable products, as shown in Figure 4 [20]. From 1988 to the present, more than 450 research papers on glycerol hydrogenolysis have been reported, according to the Scopus database. The number of publications has grown exponentially given the growing interest in the search for an alternative from a renewable source ( Figure 5). This growth also accompanies the increase in glycerol production worldwide ( Figure 3).  From 1988 to the present, more than 450 research papers on glycerol hydrogenolysis have been reported, according to the Scopus database. The number of publications has grown exponentially given the growing interest in the search for an alternative from a renewable source ( Figure 5). This growth also accompanies the increase in glycerol production worldwide ( Figure 3).
Reactions 2022, 3, FOR PEER REVIEW 5 Of all of the processes mentioned, catalytic hydrogenolysis is an interesting alternative due to the high oxygen content in the glycerol molecule (O/C = 1). In this sense, the breaking of C-O bonds of glycerol leads to the formation of glycols such as 1,2-propylene glycol (1,2-PG) and 1,3-propylene glycol (1,, and C3 alcohols such as 1-propanol (1-POH) and 2propanol (2-POH), although the breaking of C-C bonds is also present, which allows for the generation of other low molecular weight alcohols such as ethanol (EtOH) and methanol (MeOH). Of these, glycols are the most valuable products, as shown in Figure 4 [20]. From 1988 to the present, more than 450 research papers on glycerol hydrogenolysis have been reported, according to the Scopus database. The number of publications has grown exponentially given the growing interest in the search for an alternative from a renewable source ( Figure 5). This growth also accompanies the increase in glycerol production worldwide ( Figure 3).  Of the total number of papers published to date that have studied the hydrogenolysis reaction of glycerol, 75% were directed to the production of 1,2-PG, while only 21% investigated the production of 1,3-PG and the remaining 4%, the production of 1-POH, 2-POH, and EG. In addition, 92% of this research amount have studied the production of 1,3-PG in the liquid phase, while only 8% studied it in the vapor phase ( Figure 6).
Of the total number of papers published to date that have studied the hydrogenolysis reaction of glycerol, 75% were directed to the production of 1,2-PG, while only 21% investigated the production of 1,3-PG and the remaining 4%, the production of 1-POH, 2-POH, and EG. In addition, 92% of this research amount have studied the production of 1,3-PG in the liquid phase, while only 8% studied it in the vapor phase ( Figure 6). According to the data presented, clearly the hydrogenolysis of glycerol to 1,3-PG in both the liquid and vapor phase has been the least explored in the literature, and due to its projection in the market, is an attractive option to be studied.
1,3-PG has multiple applications ( Figure 7) such as the production of polyethers, polyesters, polycarbonates, and polyurethanes, which account for more than 50% of its productive application [21,22]. Currently, 30% of 1,3-PG production is used to manufacture polytrimethylene terephthalate (PTT) from terephthalic acid (Scheme 4), a polymer used in the manufacture of textile fibers and carpets due to its chemical resistance, stability, elastic recovery, and dyeing capacity. In addition, this polymer could be used as a substitute for polyethylene terephthalate (PET) and polybutene terephthalate (PBT), both from the petrochemical industry [23][24][25].  According to the data presented, clearly the hydrogenolysis of glycerol to 1,3-PG in both the liquid and vapor phase has been the least explored in the literature, and due to its projection in the market, is an attractive option to be studied.
in the liquid phase, while only 8% studied it in the vapor phase ( Figure 6). According to the data presented, clearly the hydrogenolysis of glycerol to 1,3-PG in both the liquid and vapor phase has been the least explored in the literature, and due to its projection in the market, is an attractive option to be studied.
1,3-PG has multiple applications ( Figure 7) such as the production of polyethers, polyesters, polycarbonates, and polyurethanes, which account for more than 50% of its productive application [21,22]. Currently, 30% of 1,3-PG production is used to manufacture polytrimethylene terephthalate (PTT) from terephthalic acid (Scheme 4), a polymer used in the manufacture of textile fibers and carpets due to its chemical resistance, stability, elastic recovery, and dyeing capacity. In addition, this polymer could be used as a substitute for polyethylene terephthalate (PET) and polybutene terephthalate (PBT), both from the petrochemical industry [23][24][25].  Currently, 30% of 1,3-PG production is used to manufacture polytrimethylene terephthalate (PTT) from terephthalic acid (Scheme 4), a polymer used in the manufacture of textile fibers and carpets due to its chemical resistance, stability, elastic recovery, and dyeing capacity. In addition, this polymer could be used as a substitute for polyethylene terephthalate (PET) and polybutene terephthalate (PBT), both from the petrochemical industry [23][24][25]. About 21% of the 1,3-PG market is used in the manufacture of polyurethane (PU), which is used to make elastomers and plastics. It is also used in the manufacture of cosmetics (14%), as an antifreeze fluid (9%), in the manufacture of cleaning products (8%), and as a heat exchange fluid (6%). Other applications (12%) include its use as an intermediate in the pharmaceutical industry and as an organic solvent for the manufacture of printing inks [26][27][28]. About 21% of the 1,3-PG market is used in the manufacture of polyurethane (PU), which is used to make elastomers and plastics. It is also used in the manufacture of cosmetics (14%), as an antifreeze fluid (9%), in the manufacture of cleaning products (8%), and as a heat exchange fluid (6%). Other applications (12%) include its use as an intermediate in the pharmaceutical industry and as an organic solvent for the manufacture of printing inks [26][27][28].
In the petrochemical process developed by Shell, 1,3-PG is obtained from the hydroformylation of ethylene oxide (EO) in the presence of synthesis gas (CO + H 2 ). Initially, ethylene oxide reacts with the synthesis gas, preferably at 10 MPa and 100-150 • C in the presence of a Co-Ru complex dissolved in MTBE to obtain 3-hydroxypropionaldehyde (3-HPA), which is then hydrogenated to form 1,3-PG (Scheme 5) [29]. About 21% of the 1,3-PG market is used in the manufacture of polyurethane (PU), which is used to make elastomers and plastics. It is also used in the manufacture of cosmetics (14%), as an antifreeze fluid (9%), in the manufacture of cleaning products (8%), and as a heat exchange fluid (6%). Other applications (12%) include its use as an intermediate in the pharmaceutical industry and as an organic solvent for the manufacture of printing inks [26][27][28].
In the petrochemical process developed by Shell, 1,3-PG is obtained from the hydroformylation of ethylene oxide (EO) in the presence of synthesis gas (CO + H2). Initially, ethylene oxide reacts with the synthesis gas, preferably at 10 MPa and 100-150 °C in the presence of a Co-Ru complex dissolved in MTBE to obtain 3-hydroxypropionaldehyde (3-HPA), which is then hydrogenated to form 1,3-PG (Scheme 5) [29]. In the petrochemical process developed by Degussa and later acquired by DuPont, 1,3-PG is obtained by the hydration of acrolein (Acr) and its subsequent hydrogenation. Hydration takes place in an acidic medium at 150-250 °C and hydrogenation in the presence of hydrogen and a Ni Raney catalyst at 13.5 MPa in the range of 60-140 °C (Scheme 6) [30]. Scheme 6. The formation reaction of 1,3-PG from acrolein (Acr). Although the petrochemical processes described above are the ones currently used for the production of 1,3-PG, their complexity and the toxicity of the substances involved (such as acrolein) have led us to considering other alternative routes to obtaining glycol.
On the other hand, in the fermentation process developed by Dupont, D-glucose in the presence of E. coli is used for the production of 1,3-PG [31,32]. Although this method has been able to obtain 1,3-PG with a lower content of impurities than those generated by In the petrochemical process developed by Degussa and later acquired by DuPont, 1,3-PG is obtained by the hydration of acrolein (Acr) and its subsequent hydrogenation. Hydration takes place in an acidic medium at 150-250 • C and hydrogenation in the presence of hydrogen and a Ni Raney catalyst at 13.5 MPa in the range of 60-140 • C (Scheme 6) [30]. About 21% of the 1,3-PG market is used in the manufacture of polyurethane (PU), which is used to make elastomers and plastics. It is also used in the manufacture of cosmetics (14%), as an antifreeze fluid (9%), in the manufacture of cleaning products (8%), and as a heat exchange fluid (6%). Other applications (12%) include its use as an intermediate in the pharmaceutical industry and as an organic solvent for the manufacture of printing inks [26][27][28].
In the petrochemical process developed by Shell, 1,3-PG is obtained from the hydroformylation of ethylene oxide (EO) in the presence of synthesis gas (CO + H2). Initially, ethylene oxide reacts with the synthesis gas, preferably at 10 MPa and 100-150 °C in the presence of a Co-Ru complex dissolved in MTBE to obtain 3-hydroxypropionaldehyde (3-HPA), which is then hydrogenated to form 1,3-PG (Scheme 5) [29]. In the petrochemical process developed by Degussa and later acquired by DuPont, 1,3-PG is obtained by the hydration of acrolein (Acr) and its subsequent hydrogenation. Hydration takes place in an acidic medium at 150-250 °C and hydrogenation in the presence of hydrogen and a Ni Raney catalyst at 13.5 MPa in the range of 60-140 °C (Scheme 6) [30]. Scheme 6. The formation reaction of 1,3-PG from acrolein (Acr). Although the petrochemical processes described above are the ones currently used for the production of 1,3-PG, their complexity and the toxicity of the substances involved (such as acrolein) have led us to considering other alternative routes to obtaining glycol.
On the other hand, in the fermentation process developed by Dupont, D-glucose in the presence of E. coli is used for the production of 1,3-PG [31,32]. Although this method has been able to obtain 1,3-PG with a lower content of impurities than those generated by Although the petrochemical processes described above are the ones currently used for the production of 1,3-PG, their complexity and the toxicity of the substances involved (such as acrolein) have led us to considering other alternative routes to obtaining glycol.
On the other hand, in the fermentation process developed by Dupont, D-glucose in the presence of E. coli is used for the production of 1,3-PG [31,32]. Although this method has been able to obtain 1,3-PG with a lower content of impurities than those generated by the petrochemical route, the selectivity to 1,3-PG is low due to the generation of other fermentation by-products such as EtOH, ethyl acetate, butyrate, and CO 2 . Several articles have been published using different microorganisms (Citrobacter, Clostridium, Enterobacter, Klebseilla, and Lactobacillus) to produce 1,3-PG from glycerol. The formation of secondary products such as succinic acid and acetic acid have also been observed in these studies. Moreover, since the water requirement for these processes is high, the formation of 1,3-PG proceeds very slowly, reaching high yields only at very large reaction times [33][34][35][36][37][38].
The production of 1,3-PG from the hydrogenolysis of glycerol offers a route based on the principles of green chemistry and is an alternative to the processes above-mentioned, since it would be possible to carry it out in a single reaction stage with the design of suitable catalysts. However, it has been reported that the activation energies for the removal of the primary and secondary -OH groups from the glycerol molecule are very similar (70.9 kcal/mol and 73.2 kcal/mol, respectively) and their affinity for protons is almost identical (~195 kcal/mol). These characteristics, coupled with steric hindrance, make it more difficult to cleave the secondary C-O bond than the primary one in the glycerol molecule [39]. Thus, the design of catalysts for the formation of 1,3-PG from glycerol is a real challenge.
A number of review papers have so far been published on the subject, focusing their study on the properties of catalysts and their catalytic performance [40][41][42][43][44][45].
The first one dates back to 2013, published by Feng et al. who, after reviewing the literature data, concluded that catalysts based on Pt, Rh, or Cu in the presence of W chemical species were promising for the production of 1,3-PG, but the relationship between the metal phase and W species was unknown until that time [40]. In 2014, Nakagawa et al. published a review emphasizing the difficulty in selectively obtaining 1,3-PG in glycerol hydrogenolysis and punctuating the fact that noble metal-based catalysts such as Pt and Ir modified with W or Re are especially effective in the production of glycol. The authors elucidated that the hydrogenolysis of glycerol to 1,3-PG is a catalytic structure sensitive reaction and the catalyst composition must be cautiously designed to obtain high yields [41].
In 2020, Wang et al. reported the performance of different Pt-W catalysts and the properties of the catalysts that influence the activity and selectivity toward the formation of 1,3-PG [42]. In 2021, Wu et al. complemented the review by focusing in detail on the relationship between W species and the metal phase to selectively produce 1,3-PG from glycerol [43].
A review article has also reported on the current 1,3-PG market and its growth prospects in future years, considering the current chemical production routes and the development of the latest heterogeneous catalysts aiming at new economically competitive industrial processes [44].
Recently, Zhang et al. published an article about the key factors for the design of active and selective catalysts toward the production of 1,3-PG. They performed a critical review on the electronic effects and steric hindrances at the metal-acid interfaces of catalytic surfaces. Their results provide important insights for manipulating the spatial structure of a given catalyst to produce 1,3-PG selectively [45].
The present review paper proposes to review the articles reporting the hydrogenolysis of glycerol for the production of 1,3-PG focusing on engineering aspects, describing and comparing the catalytic systems used in liquid phase batch reactors and continuous flow reactors, both in the liquid and vapor phase. As an introduction, the main reaction mechanisms and the performance of the main catalysts used are presented, classifying them according to their metallic phase and support. Since the reactor configuration has an impact on the product distribution and catalytic stability, the effect of the operating conditions in the reactors was analyzed and concluded with a review of the stability of the catalytic systems. The novel aspect of this review, in contrast to previously reported works, is the incorporation of a section dedicated to engineering aspects and the effect of reactor parameters, which will serve as a tool to design new effective catalysts and determine their ideal operating conditions to maximize 1,3-PG production.

Mechanisms of Glycerol Hydrogenolysis for the Production of 1,3-PG
Hydrogenolysis is a type of reduction reaction in which the breaking of C-O and C-C bonds occurs with the addition of hydrogen, often using catalysts of different natures (Scheme 7). While the breaking of C-C bonds decreases the length of a molecule and fragments it into smaller molecules, the breaking of C-O bonds decreases the oxygen content in the starting molecule, maintaining its total number of carbon atoms.

Mechanisms of Glycerol Hydrogenolysis for the Production of 1,3-PG
Hydrogenolysis is a type of reduction reaction in which the breaking of C-O and C-C bonds occurs with the addition of hydrogen, often using catalysts of different natures (Scheme 7). While the breaking of C-C bonds decreases the length of a molecule and fragments it into smaller molecules, the breaking of C-O bonds decreases the oxygen content in the starting molecule, maintaining its total number of carbon atoms. In particular, the hydrogenolysis of glycerol can lead to the formation of 1,2-PG, 1,3-PG, 1-POH, 2-POH, and propane by the C-O bond cleavage, while the C-C bond cleavage can lead to the formation of EG, MeOH, EtOH, methane, and ethane (Scheme 8) [9]. The mechanism of the hydrogenolysis of glycerol varies depending on the operating conditions and catalysts employed.
The first of the reports was published by Montassier et al., who studied the selective hydrogenolysis of glycerol solutions at very low concentrations (1-4 wt.%) using a Ru/C catalyst, moderate temperatures (180-260 °C), and high pressures (3-4 MPa) [46]. Their results showed the presence of glyceraldehyde (GLA), 2-hydroxyacrolein (2-HA), and 1,2-PG, suggesting that the reaction proceeded via a mechanism involving the In particular, the hydrogenolysis of glycerol can lead to the formation of 1,2-PG, 1,3-PG, 1-POH, 2-POH, and propane by the C-O bond cleavage, while the C-C bond cleavage can lead to the formation of EG, MeOH, EtOH, methane, and ethane (Scheme 8) [9]. C bonds occurs with the addition of hydrogen, often using catalysts of different natures (Scheme 7). While the breaking of C-C bonds decreases the length of a molecule and fragments it into smaller molecules, the breaking of C-O bonds decreases the oxygen content in the starting molecule, maintaining its total number of carbon atoms. In particular, the hydrogenolysis of glycerol can lead to the formation of 1,2-PG, 1,3-PG, 1-POH, 2-POH, and propane by the C-O bond cleavage, while the C-C bond cleavage can lead to the formation of EG, MeOH, EtOH, methane, and ethane (Scheme 8) [9]. The mechanism of the hydrogenolysis of glycerol varies depending on the operating conditions and catalysts employed.
The first of the reports was published by Montassier et al., who studied the selective hydrogenolysis of glycerol solutions at very low concentrations (1-4 wt.%) using a Ru/C catalyst, moderate temperatures (180-260 °C), and high pressures (3-4 MPa) [46]. Their results showed the presence of glyceraldehyde (GLA), 2-hydroxyacrolein (2-HA), and 1,2-PG, suggesting that the reaction proceeded via a mechanism involving the C C The mechanism of the hydrogenolysis of glycerol varies depending on the operating conditions and catalysts employed.
The first of the reports was published by Montassier et al., who studied the selective hydrogenolysis of glycerol solutions at very low concentrations (1-4 wt.%) using a Ru/C catalyst, moderate temperatures (180-260 • C), and high pressures (3-4 MPa) [46]. Their results showed the presence of glyceraldehyde (GLA), 2-hydroxyacrolein (2-HA), and 1,2-PG, suggesting that the reaction proceeded via a mechanism involving the dehydrogenation of glycerol to produce GLA at the metal sites of the catalyst, followed by the dehydration of GLA to produce 2-HA in the presence of a slightly basic aqueous medium, and finally, the hydrogenation of 2-HA to produce 1,2-PG at the metal sites of the catalyst (Scheme 9). Both dehydrogenation and hydrogenation processes take place on the metal sites of the catalyst (Ru particles), while dehydration process occurs because of the basic medium.
dehydrogenation of glycerol to produce GLA at the metal sites of the catalyst, followed by the dehydration of GLA to produce 2-HA in the presence of a slightly basic aqueous medium, and finally, the hydrogenation of 2-HA to produce 1,2-PG at the metal sites of the catalyst (Scheme 9). Both dehydrogenation and hydrogenation processes take place on the metal sites of the catalyst (Ru particles), while dehydration process occurs because of the basic medium. The dehydrogenation step of glycerol to GLA is thermodynamically disadvantaged due to the presence of H2, especially when the reaction is carried out at high H2 pressures. In these cases, GLA can be rapidly hydrogenated back to glycerol. Therefore, to favor the Montassier mechanism, the dehydration step must be very fast, and this can occur in the presence of bases in the reaction medium or basic catalysts.
Montassier et al. proposed two different routes for the formation of side products from C-C bond cleavage reactions.
When a Ru catalyst is employed, they proposed that the formation of EG and MeOH comes from a retro-aldol reaction, as indicated in Scheme 10a. Subsequently, MeOH could be further hydrogenated to generate CH4 and H2O. The higher tendency for C-C bond cleavage allows for low molecular weight hydrocarbons such as methane using the Ru catalyst to be obtained. The dehydrogenation step of glycerol to GLA is thermodynamically disadvantaged due to the presence of H 2 , especially when the reaction is carried out at high H 2 pressures. In these cases, GLA can be rapidly hydrogenated back to glycerol. Therefore, to favor the Montassier mechanism, the dehydration step must be very fast, and this can occur in the presence of bases in the reaction medium or basic catalysts.
Montassier et al. proposed two different routes for the formation of side products from C-C bond cleavage reactions.
When a Ru catalyst is employed, they proposed that the formation of EG and MeOH comes from a retro-aldol reaction, as indicated in Scheme 10a. Subsequently, MeOH could be further hydrogenated to generate CH 4 and H 2 O. The higher tendency for C-C bond cleavage allows for low molecular weight hydrocarbons such as methane using the Ru catalyst to be obtained.
dehydrogenation of glycerol to produce GLA at the metal sites of the catalyst, followed by the dehydration of GLA to produce 2-HA in the presence of a slightly basic aqueous medium, and finally, the hydrogenation of 2-HA to produce 1,2-PG at the metal sites of the catalyst (Scheme 9). Both dehydrogenation and hydrogenation processes take place on the metal sites of the catalyst (Ru particles), while dehydration process occurs because of the basic medium. The dehydrogenation step of glycerol to GLA is thermodynamically disadvantaged due to the presence of H2, especially when the reaction is carried out at high H2 pressures. In these cases, GLA can be rapidly hydrogenated back to glycerol. Therefore, to favor the Montassier mechanism, the dehydration step must be very fast, and this can occur in the presence of bases in the reaction medium or basic catalysts.
Montassier et al. proposed two different routes for the formation of side products from C-C bond cleavage reactions.
When a Ru catalyst is employed, they proposed that the formation of EG and MeOH comes from a retro-aldol reaction, as indicated in Scheme 10a. Subsequently, MeOH could be further hydrogenated to generate CH4 and H2O. The higher tendency for C-C bond cleavage allows for low molecular weight hydrocarbons such as methane using the Ru catalyst to be obtained.  When a Cu catalyst is used, the authors proposed that the formation of EG comes from a retro-Claisen reaction, as indicated in Scheme 10b. In this case, the ability to cleave C-O bonds with respect to C-C bonds is higher for the Cu catalyst, leading to a lower selectivity to EG and MeOH.
These first results, although not focused on the mechanism of obtaining 1,3-PG, show that the selective conversion of glycerol to propanediols requires bifunctional catalysts that promote both dehydration and hydrogenation reactions.
In this section, each of the reported reaction mechanisms for the formation of 1,3-PG are presented in order of chronological occurrence.

Mechanism of Dehydration-Hydrogenation
Dasari et al. carried out tests in a liquid phase batch reactor at 200 • C and 1.4 MPa H 2 , using a Cu/Cr 2 O 4 catalyst with 80 wt.% analytical glycerol solutions and 24 h of reaction time. From their results, they proposed the formation of 1,2-PG from acetol (AcOH) as a reaction intermediate [47]. In this mechanism, glycerol is first dehydrated to AcOH, which is then hydrogenated to form 1,2-PG (Scheme 11). The presence of acid sites in Cu/Cr 2 O 4 favor the dehydration step to form AcOH from glycerol, and the presence of the metallic phase of Cu makes the hydrogenation step possible to occur.
C-O bonds with respect to C-C bonds is higher for the Cu catalyst, leading to a lower selectivity to EG and MeOH.
These first results, although not focused on the mechanism of obtaining 1,3-PG, show that the selective conversion of glycerol to propanediols requires bifunctional catalysts that promote both dehydration and hydrogenation reactions.
In this section, each of the reported reaction mechanisms for the formation of 1,3-PG are presented in order of chronological occurrence.

Mechanism of Dehydration-Hydrogenation
Dasari et al. carried out tests in a liquid phase batch reactor at 200 °C and 1.4 MPa H2, using a Cu/Cr2O4 catalyst with 80 wt.% analytical glycerol solutions and 24 h of reaction time. From their results, they proposed the formation of 1,2-PG from acetol (AcOH) as a reaction intermediate [47]. In this mechanism, glycerol is first dehydrated to AcOH, which is then hydrogenated to form 1,2-PG (Scheme 11). The presence of acid sites in Cu/Cr2O4 favor the dehydration step to form AcOH from glycerol, and the presence of the metallic phase of Cu makes the hydrogenation step possible to occur. From the thermodynamic point of view, it has been reported that the formation of 1,2-PG from AcOH is a reversible process in which the direct reaction rate (AcOH → 1,2-PG) is higher than the reverse reaction rate (1,2-PG → AcOH), so it is expected that in the presence of H2, the route of transformation of AcOH to 1,2-PG is the most favorable one [48]. Recently, Yfanti et al., making use of the in situ IR spectroscopy technique, reported that the formation of 1,2-PG from glycerol occurs via AcOH, and such formation is favored at a higher availability of hydrogen in the reaction medium [49].
Although the reaction mechanism proposed by Dasari et al. includes the formation of neither 1,3-PG nor the 1-POH and 2-POH propanols, their work was important as it was a breakthrough in the study of the reaction mechanism of hydrogenolysis [50].
In 2006, Miyazawa et al. proved that the dehydration of glycerol to AcOH was catalyzed by the use of acidic materials (H2SO4, HCl, and Amberlyst resin), and the subsequent hydrogenation of AcOH over the metal sites of the catalyst (Ru, Pt, Rh, Pd) led to the formation of 1,2-PG. Additionally, the results of their tests showed the presence of 1,3-PG for the first time, and a route to obtain it in two stages was proposed by analogy to the formation of 1,2-PG. According to this route, glycerol is first dehydrated to 3-hydroxypropionaldehyde (3-HPA), which is then hydrogenated to generate 1,3-PG (Scheme 12) [51].
Scheme 11. The reaction mechanism proposed by Dasari et al. for the formation of 1,2-PG.
From the thermodynamic point of view, it has been reported that the formation of 1,2-PG from AcOH is a reversible process in which the direct reaction rate (AcOH → 1,2-PG) is higher than the reverse reaction rate (1,2-PG → AcOH), so it is expected that in the presence of H 2 , the route of transformation of AcOH to 1,2-PG is the most favorable one [48]. Recently, Yfanti et al., making use of the in situ IR spectroscopy technique, reported that the formation of 1,2-PG from glycerol occurs via AcOH, and such formation is favored at a higher availability of hydrogen in the reaction medium [49].
Although the reaction mechanism proposed by Dasari et al. includes the formation of neither 1,3-PG nor the 1-POH and 2-POH propanols, their work was important as it was a breakthrough in the study of the reaction mechanism of hydrogenolysis [50].
In 2006, Miyazawa et al. proved that the dehydration of glycerol to AcOH was catalyzed by the use of acidic materials (H 2 SO 4 , HCl, and Amberlyst resin), and the subsequent hydrogenation of AcOH over the metal sites of the catalyst (Ru, Pt, Rh, Pd) led to the formation of 1,2-PG. Additionally, the results of their tests showed the presence of 1,3-PG for the first time, and a route to obtain it in two stages was proposed by analogy to the formation of 1,2-PG. According to this route, glycerol is first dehydrated to 3-hydroxypropionaldehyde (3-HPA), which is then hydrogenated to generate 1,3-PG (Scheme 12) [51]. From a thermodynamic point of view, AcOH is a more stable intermediate than 3-HPA, being identified in most scientific reports [40], and the formation of 1,2-PG predominates over the formation of 1,3-PG [52].
Hydrogenation can lead to secondary products such as 1-POH and 2-POH. Miyazawa et al. also proposed the formation of EG and MeOH and the subsequent hydrogenation of EG to produce EtOH (Scheme 13) [51]. From a thermodynamic point of view, AcOH is a more stable intermediate than 3-HPA, being identified in most scientific reports [40], and the formation of 1,2-PG predominates over the formation of 1,3-PG [52].
Hydrogenation can lead to secondary products such as 1-POH and 2-POH. Miyazawa et al. also proposed the formation of EG and MeOH and the subsequent hydrogenation of EG to produce EtOH (Scheme 13) [51].
From a thermodynamic point of view, AcOH is a more stable intermediate than 3-HPA, being identified in most scientific reports [40], and the formation of 1,2-PG predominates over the formation of 1,3-PG [52].
Hydrogenation can lead to secondary products such as 1-POH and 2-POH. Miyazawa et al. also proposed the formation of EG and MeOH and the subsequent hydrogenation of EG to produce EtOH (Scheme 13) [51]. The formation of propanols from 1,2-PG and 1,3-PG was reported by Perosa et al. who, using a Ni-Raney catalyst, detected the formation of secondary products. In their reaction scheme, they established the formation of 1-propanol (1-POH) and 2-propanol (2-POH) from 1,3-PG and 1,2-PG, respectively, which in turn can be further hydrogenated to form propane (Scheme 14) [54]. The presence of 1-POH, 2-POH and propane could be explained because of a certain tendency of Ni-Raney to cleave C-O bonds.  Maglinao et al. calculated the vapor phase equilibrium K p constants for several of the reactions involved in this hydrogenolysis mechanism. Regarding the first dehydration step, the reactions were reported to be thermodynamically favorable, both the dehydration of glycerol to AcOH (K p (190 • C)~7.3 × 10 9 ) and the dehydration of glycerol to 3-HPA (K p (190 • C)~8.2 × 10 7 ). These values confirm that the formation of AcOH is more favorable than the formation of 3-HPA. With respect to the second stage, the hydrogenation of the formed intermediates is limited by thermodynamic equilibrium, both the hydrogenation of AcOH to 1,2-PG (K p (190 • C)~10 −1 ), and the hydrogenation of 3-HPA to 1,3-PG (K p (190 • C)~2.6 × 10 −2 ) [53].
The formation of propanols from 1,2-PG and 1,3-PG was reported by Perosa et al. who, using a Ni-Raney catalyst, detected the formation of secondary products. In their reaction scheme, they established the formation of 1-propanol (1-POH) and 2-propanol (2-POH) from 1,3-PG and 1,2-PG, respectively, which in turn can be further hydrogenated to form propane (Scheme 14) [54]. The presence of 1-POH, 2-POH and propane could be explained because of a certain tendency of Ni-Raney to cleave C-O bonds.   Furikado et al. found that the formation of propanols followed a different reaction pathway for Rh/C catalysts with respect to the Ru/C catalysts. In the case of Rh/C, 2-POH is formed from 1,2-PG, whereas for Ru/C, only 1-POH is present, which could come from 1,2-PG [55]. Kunosoki et al. suggested the formation of 1-POH by both 1,2-PG and 1,3-PG hydrogenation, and additionally proposed the formation of MeOH and EtOH from both glycols (Scheme 15) [56]. Scheme 14. The process of obtaining 1-POH, 2-POH, and propane from glycols, according to Perosa et al. Furikado et al. found that the formation of propanols followed a different reaction pathway for Rh/C catalysts with respect to the Ru/C catalysts. In the case of Rh/C, 2-POH is formed from 1,2-PG, whereas for Ru/C, only 1-POH is present, which could come from 1,2-PG [55]. Kunosoki et al. suggested the formation of 1-POH by both 1,2-PG and 1,3-PG hydrogenation, and additionally proposed the formation of MeOH and EtOH from both glycols (Scheme 15) [56]. With respect to EtOH formation, Magliano et al. indicated that although it could occur by the hydrogenolysis of either 1,2-PG or AcOH, the former thermodynamically predominated over the latter. However, as both equilibrium constant values are high, the design and properties of the catalyst determine which of the two reactions has an impact on EtOH formation [48]. With respect to EtOH formation, Magliano et al. indicated that although it could occur by the hydrogenolysis of either 1,2-PG or AcOH, the former thermodynamically predominated over the latter. However, as both equilibrium constant values are high, the design and properties of the catalyst determine which of the two reactions has an impact on EtOH formation [48].

Mechanism of Chelation-Hydrogenation
Chaminand et al. employed Cu, Pd, and Rh catalysts supported on C, ZnO, and Al 2 O 3 to carry out the hydrogenolysis reaction at 180 • C and 8 MPa H 2 . Of the catalysts studied, Rh/C, promoted by H 2 WO 4 , allowed 1,2-PG and 1,3-PG to be obtained. By replacing water with different solvents (sulfolane and dioxane), they were able to increase the glycerol conversion and make the reaction more selective to each of these glycols. Based on their results, these authors proposed a mechanism with three probable routes for the formation of glycols. One of them consisted of a chelation step on the metal sites and subsequent hydrogenation in the presence of H 2 (Scheme 16) [57].
The chelation of glycerol on the metal sites of the catalyst (M) and its subsequent hydrogenation suggest that the formation of 1,3-PG is favored over the formation of 1,2-PG due to the stability of the intermediate formed. However, this was the only work that proposed chelation as an intermediate process in the formation of 1,3-PG.
to carry out the hydrogenolysis reaction at 180 °C and 8 MPa H2. Of the catalysts studied, Rh/C, promoted by H2WO4, allowed 1,2-PG and 1,3-PG to be obtained. By replacing water with different solvents (sulfolane and dioxane), they were able to increase the glycerol conversion and make the reaction more selective to each of these glycols. Based on their results, these authors proposed a mechanism with three probable routes for the formation of glycols. One of them consisted of a chelation step on the metal sites and subsequent hydrogenation in the presence of H2 (Scheme 16) [57].

Direct Mechanism of Hydrogenolysis
Some articles have reported the hydrogenolysis of glycerol through a direct mechanism.
Tomishige et al. reported Ir-MOx/SiO2 and Rh-MOx/SiO2 (M = Re, Mo, V) catalysts that were studied in the liquid phase reaction, in the range of 100-180 °C and pressures between 2 and 8 MPa of H2, employing aqueous glycerol solutions between 20 and 100 wt.%. Taking as reference the Ir-ReOx/SiO2 and Rh-ReOx/SiO2 catalysts, they formulated a direct mechanism of hydrogenolysis based on the formation of hydroxypropoxides on the catalytic surface, leading to the formation of 1,2-PG or 1,3-PG depending on the intermediate formed.

Direct Mechanism of Hydrogenolysis
Some articles have reported the hydrogenolysis of glycerol through a direct mechanism. Tomishige et al. reported Ir-MO x /SiO 2 and Rh-MO x /SiO 2 (M = Re, Mo, V) catalysts that were studied in the liquid phase reaction, in the range of 100-180 • C and pressures between 2 and 8 MPa of H 2 , employing aqueous glycerol solutions between 20 and 100 wt.%. Taking as reference the Ir-ReO x /SiO 2 and Rh-ReO x /SiO 2 catalysts, they formulated a direct mechanism of hydrogenolysis based on the formation of hydroxypropoxides on the catalytic surface, leading to the formation of 1,2-PG or 1,3-PG depending on the intermediate formed.
According to this mechanism, as represented in Scheme 17, glycerol adsorbs on the surface sites of the ReO x species forming 1,3-dihydroxypropoxide or 2,3-dihydroxypropoxide (1st step). Then, H 2 , dissociatively adsorbed on the metal surface (Ir or Rh), attacks the tertiary or secondary carbon of the intermediate species, forming 1-hydroxypropoxide or 3-hydroxypropoxide, respectively (2nd step). Finally, these species are hydrolyzed to give rise to the formation of 1,2-PG or 1,3-PG, respectively (3rd step) [58][59][60]. Guan et al. employed Ir-ReOx/ZrO2 catalysts and DFT calculations that allowed them to conclude that the formation of 1,3-PG proceeded by a direct reaction mechanism via alkoxide species formed on the ReOx particles. In this regard, the results showed that the formation of terminal alkoxide species (2,3-dihydroxypropoxide) prevailed over the formation of secondary alkoxide species (1,3-dihydroxypropoxide), whereby the formation of 1,3-PG was favored over the formation of 1,2-PG [61]. These results agree with the catalytic performance of Ir-Re/SiO2 reported by Tomishige's group [58][59][60].
Another direct mechanism of glycerol hydrogenolysis was proposed for the Pt/WO3/ZrO2 [62], Pt/WOx [63], Pt/WOx/AlOOH [64], Pt/WOx/Al2O3 [65,66], and Pt/WOx/SBA-15 [67] systems. It has been proposed that on these catalysts, the formation of glycols occurs through an ionic mechanism involving proton and hydride transfer steps on the catalytic surface [62]. Although the mechanism presents variations depending on the support employed, the formation of glycols (1,2-PG or 1,3-PG) starts with the adsorp-  Guan et al. employed Ir-ReO x /ZrO 2 catalysts and DFT calculations that allowed them to conclude that the formation of 1,3-PG proceeded by a direct reaction mechanism via alkoxide species formed on the ReO x particles. In this regard, the results showed that the formation of terminal alkoxide species (2,3-dihydroxypropoxide) prevailed over the formation of secondary alkoxide species (1,3-dihydroxypropoxide), whereby the formation of 1,3-PG was favored over the formation of 1,2-PG [61]. These results agree with the catalytic performance of Ir-Re/SiO 2 reported by Tomishige's group [58][59][60].
Another direct mechanism of glycerol hydrogenolysis was proposed for the Pt/WO 3 / ZrO 2 [62], Pt/WO x [63], Pt/WO x /AlOOH [64], Pt/WO x /Al 2 O 3 [65,66], and Pt/WO x / SBA-15 [67] systems. It has been proposed that on these catalysts, the formation of glycols occurs through an ionic mechanism involving proton and hydride transfer steps on the catalytic surface [62]. Although the mechanism presents variations depending on the support employed, the formation of glycols (1,2-PG or 1,3-PG) starts with the adsorption of glycerol on the support and the adsorption of H 2 through a heterolytic cleavage into H + and H − on the metal sites of the active phase (1st step). Next, protonation and dehydration of the adsorbed glycerol occurs to form a primary (I) or secondary (II) oxocarbenium type intermediate, depending on the attack of the H + on the primary or secondary -OH group of the adsorbed glycerol (2nd step). Finally, the H − ions are transferred to the oxocarbenium (I) or (II) ions to form 1,2-PG or 1,3-PG, respectively (3rd step) (Scheme 18).
Guan et al. employed Ir-ReOx/ZrO2 catalysts and DFT calculations that allowed them to conclude that the formation of 1,3-PG proceeded by a direct reaction mechanism via alkoxide species formed on the ReOx particles. In this regard, the results showed that the formation of terminal alkoxide species (2,3-dihydroxypropoxide) prevailed over the formation of secondary alkoxide species (1,3-dihydroxypropoxide), whereby the formation of 1,3-PG was favored over the formation of 1,2-PG [61]. These results agree with the catalytic performance of Ir-Re/SiO2 reported by Tomishige's group [58][59][60].
Another direct mechanism of glycerol hydrogenolysis was proposed for the Pt/WO3/ZrO2 [62], Pt/WOx [63], Pt/WOx/AlOOH [64], Pt/WOx/Al2O3 [65,66], and Pt/WOx/SBA-15 [67] systems. It has been proposed that on these catalysts, the formation of glycols occurs through an ionic mechanism involving proton and hydride transfer steps on the catalytic surface [62]. Although the mechanism presents variations depending on the support employed, the formation of glycols (1,2-PG or 1,3-PG) starts with the adsorption of glycerol on the support and the adsorption of H2 through a heterolytic cleavage into H+ and Hon the metal sites of the active phase (1 st step). Next, protonation and dehydration of the adsorbed glycerol occurs to form a primary (I) or secondary (II) oxocarbenium type intermediate, depending on the attack of the H + on the primary or secondary -OH group of the adsorbed glycerol (2 nd step). Finally, the Hions are transferred to the oxocarbenium (I) or (II) ions to form 1,2-PG or 1,3-PG, respectively (3 rd step) (Scheme 18). It has been reported that the choice of support (S) in the Pt/WO x /S catalysts plays a fundamental role in the reaction mechanism since it disperses and stabilizes the Pt metal particles and WO x species [63]. The presence of WO x species is fundamental because they allow for a better dispersion of Pt, which facilitates the heterolytic cleavage of H 2 molecules into H + and H − [63,65], and on the other hand, they provide sites for the adsorption of glycerol and the stabilization of intermediate carbocations [66]. According to recent reports, WO x species have the character of Brønsted acids and favor the formation of 1,3-PG [67].
Theoretical studies based on DFT calculations have shown that oxocarbenium-type intermediates are relatively stable and that the formation of other enol-type intermediates from oxocarbeniums is therefore unlikely [68]. The results showed that the hydrogenolysis of glycerol occurred via a bifunctional mechanism involving the dehydration of glycerol to AcOH and glycidol on the surface of ZnO particles and the hydrogenation of these intermediates on Cu particles. Both glycerol conversion and selectivity to 1,2-PG depend on the sizes of the ZnO and Cu particles. The authors proposed a new pathway for the formation of 1,2-PG, which is presented in Scheme 19 [69].

Mechanism of Etherification-Hydrogenation
diate. They employed Cu-ZnO catalysts, prepared by the coprecipitation method, at 180-240 °C and 4.2 MPa H2. The results showed that the hydrogenolysis of glycerol occurred via a bifunctional mechanism involving the dehydration of glycerol to AcOH and glycidol on the surface of ZnO particles and the hydrogenation of these intermediates on Cu particles. Both glycerol conversion and selectivity to 1,2-PG depend on the sizes of the ZnO and Cu particles. The authors proposed a new pathway for the formation of 1,2-PG, which is presented in Scheme 19 [69].    on the surface of ZnO particles and the hydrogenation of these intermediates on Cu particles. Both glycerol conversion and selectivity to 1,2-PG depend on the sizes of the ZnO and Cu particles. The authors proposed a new pathway for the formation of 1,2-PG, which is presented in Scheme 19 [69].

Catalysts for 1,3-PG Production
This section presents the most relevant results of the activity and catalytic properties of the catalysts used for the production of 1,3-PG, differentiating them according to the metal phase and the type of reactor, both in the liquid and vapor phase.

Pd, Ru, and Rh Catalysts
The first results with a certain tendency to form 1,3-PG were obtained using homogeneous phase catalysts based on Pd, Ru, and Rh as the active phase. Table 1 summarizes the results obtained in the liquid phase batch reactors. m gly /m Me -glycerol/metal mass ratio; NR-not reported; X-glycerol conversion; S 1,3-PG -selectivity to 1,3-PG; Y 1,3-PG -yield to 1,3-PG.
One of the first references dates back to a patent developed by Che, who reported that by employing a catalyst composed of Rh and H 2 WO 4 in the presence of 1-methyl-2pyrrolidinone as the solvent, it is possible to achieve a glycerol conversion of 47% with a selectivity to 1,3-PG of 43% at 200 • C and 31.6 MPa of a CO/H 2 (1:2) mixture after 24 h of reaction [71].
Bullock et al. reported the use of homogeneous Ru catalysts in a sulfolane medium for the production of 1,3-PG. Under moderate temperature and pressure conditions, 110 • C and 5.2 MPa H 2 , they obtained very low yields to 1,3-PG (~3%) [73].
The difficulty in separating the homogeneous catalyst and recovering it from the reaction product mixture has led to the consideration of heterogeneous phase catalytic systems.
Given the superior activity of Rh-based homogeneous catalysts, the first supported catalysts used Rh as the active phase. Table 2 summarizes the main results obtained with these catalysts used in liquid-phase batch reactors.   [56]. With respect to the hydrogenating activity, they found the following order of activity: Ru/C > Pt/C > Rh/C > Pd/C Following the work of Chaminand et al., the results by Kunosoki et al. confirmed that Rh is an interesting metal for obtaining 1,3-PG due to the selectivity values obtained with respect to Ru, Pt, and Pd. By also adding H 2 WO 4 to the reaction medium in the presence of Ru/C and Rh/C, they obtained a marked increase in the selectivity to 1,3-PG in the presence of Rh/C, suggesting the existence of a Rh-W bond that favors the production of glycol [56]. Surprisingly, there have been no studies following up on this idea.
Furikado et al. tested Rh/SiO 2 catalysts that were equally selective to 1,3-PG as Rh/C catalysts. In the presence of an Amberlyst resin in the reaction medium, the activity levels improved, but there was no substantial change in selectivity to 1,3-PG [55]. Other Rhbased systems on heteropolyacids, Cs 2.5 H 0.5 [PW 12 O 40 ], also showed some tendency to form 1,3-PG [74]. However, in all cases, the yield to 1,3-PG was very low (<4%).
Higher, but still low yields (12%) were obtained by employing a ReO x -modified Rh/SiO 2 catalyst. The results indicated that the acidic properties of the ReO x particles favored the formation of the reaction intermediate, leading to the formation of 1,3-PG following the direct mechanism of hydrogenolysis [75].
These results, although showing low yields to 1,3-PG, proved that the presence of an active hydrogenating phase and acidic chemical species such as H 2 WO 4 or ReO x particles is promising for the development of improved catalysts for the production of 1,3-PG.

Ir Catalysts
Ir catalysts and their bimetallic combinations with Ni, Co, Re, and Ru have been reported with some activity for the formation of 1,3-PG. Reports have employed these catalysts in the liquid phase using batch reactors (Table 3). m gly /m c -glycerol/catalyst mass ratio; X-glycerol conversion; S 1,3-PG -selectivity to 1,3-PG; Y 1,3-PG -yield to 1,3-PG.
Low yields (<1%) were obtained by employing Ir/γ-Al 2 O 3 catalysts and their modification with Ni, Ir-Ni/γ-Al 2 O 3 , even though the presence of the metal produced an increase in glycerol conversion. For these catalysts, the species responsible for the activity were IrO 2 and IrCl x coming from the decomposition of the precursor H 2 IrCl 6 .xH 2 O [76]. In contrast, Co-Ir nanorods (Co:Ir = 100) were shown to be more active toward the formation of 1,3-PG, essentially due to the morphology of these structures and the exposure of their {10-10} crystalline planes [77].
Ir-Re based catalytic systems such as Ir-ReO x /SiO 2 showed even higher yields due to a promoting effect by Re [59][60][61]78,79,86,87]. The presence of ReO x particles results in the formation of 2,3-dihydroxypropoxide from glycerol. Then, gaseous hydrogen, activated on the metal sites of Ir under the hydride form, attacks the 2-position of 2,3-dihydroxypropoxide to form 3-hydroxypropoxide, which finally hydrolyzes to lead to the formation of 1,3-PG ( Figure 8) [59].
Modifying the Ir/SiO 2 catalyst with other oxides based on W, Mo, V, Cr, Mn, and Ag, Amada et al. found that the selectivity to 1,3-PG varied in the following order: Their study determined that the cleavage of the C-O bond occurred at the interface between the metal surface of the Ir and the ReO x particles, where the substrate was adsorbed [60]. Unlike other promoters, ReO x allows for the formation of a terminal alkoxide, leading to the formation of 1,3-PG instead of a secondary alkoxide that would favor the formation of 1,2-PG [61].
Ir-Re based catalytic systems such as Ir-ReOx/SiO2 showed even higher yield a promoting effect by Re [59][60][61]78,79,86,87]. The presence of ReOx particles resu formation of 2,3-dihydroxypropoxide from glycerol. Then, gaseous hydrogen, a on the metal sites of Ir under the hydride form, attacks the 2-position of 2,3-dih propoxide to form 3-hydroxypropoxide, which finally hydrolyzes to lead to the fo of 1,3-PG (Figure 8) [59]. Their study determined that the cleavage of the C-O bond occurred at the between the metal surface of the Ir and the ReOx particles, where the substrate sorbed [60]. Unlike other promoters, ReOx allows for the formation of a terminal a leading to the formation of 1,3-PG instead of a secondary alkoxide that would f formation of 1,2-PG [61].
In their study, they further proved that by keeping the Re promoter but chan metal phase to Rh, Ru, and Pt, the order of activity was as follows: Finally, they concluded that the performance of Ir-ReOx/SiO2 was superio presented by the Ir/SiO2 and ReOx/SiO2 monometallic catalysts [60].
To understand a little more about their promoter effect, Gilkey et al. studied t tural properties of ReOx particles on the Ir-ReOx/SiO2 catalysts. During prepara catalyst precursor is in the form of IrO2-ReOx/SiO2 with Ir having a +4 oxidation s is reduced to Ir 0 after reduction. The ReOx species, on the other hand, are partially and remain with an average oxidation state of +2.6 after reduction [86]. Ama demonstrated that, once reduced, the surface of such catalysts was covered by m particles forming cube-octahedrons covered by ReOx particles. The presence o proved the reducibility of the catalyst due to the interaction between Re and IrO In their study, they further proved that by keeping the Re promoter but changing the metal phase to Rh, Ru, and Pt, the order of activity was as follows: Ir > Pt > Rh > Ru > Pd Finally, they concluded that the performance of Ir-ReO x /SiO 2 was superior to that presented by the Ir/SiO 2 and ReO x /SiO 2 monometallic catalysts [60].
To understand a little more about their promoter effect, Gilkey et al. studied the structural properties of ReO x particles on the Ir-ReO x /SiO 2 catalysts. During preparation, the catalyst precursor is in the form of IrO 2 -ReO x /SiO 2 with Ir having a +4 oxidation state that is reduced to Ir 0 after reduction. The ReO x species, on the other hand, are partially reduced and remain with an average oxidation state of +2.6 after reduction [86]. Amada et al. demonstrated that, once reduced, the surface of such catalysts was covered by metallic Ir particles forming cube-octahedrons covered by ReO x particles. The presence of Re improved the reducibility of the catalyst due to the interaction between Re and IrO 2 during the calcination process. In addition, the performance of the Ir-ReO x /SiO 2 catalysts was attributed to the active sites formed by metal bonds between metallic Ir and ReO x species (Figure 9) [87].
In the presence of another metal such as Ru, the Ir-ReO x /SiO 2 catalysts showed enhanced activity levels toward the formation of 1,3-PG. The presence of Ru at a given concentration dissociates H 2 molecules and transfers H species via a spillover mechanism toward the Ir-ReO x particles. These species form protons and facilitate the formation of 1,3-PG ( Figure 10) [80]. Surprisingly, Ru-Re/SiO 2 catalysts were found to be less active and selective than the Ir-ReO x /SiO 2 catalysts, reinforcing the idea that Ru acts as a promoter in the hydrogenolysis reaction [88].
Reactions 2022, 3, FOR PEER REVIEW 20 the calcination process. In addition, the performance of the Ir-ReOx/SiO2 catalysts was attributed to the active sites formed by metal bonds between metallic Ir and ReOx species (Figure 9) [87]. In the presence of another metal such as Ru, the Ir-ReOx/SiO2 catalysts showed enhanced activity levels toward the formation of 1,3-PG. The presence of Ru at a given concentration dissociates H2 molecules and transfers H species via a spillover mechanism toward the Ir-ReOx particles. These species form protons and facilitate the formation of 1,3-PG ( Figure 10) [80]. Surprisingly, Ru-Re/SiO2 catalysts were found to be less active and selective than the Ir-ReOx/SiO2 catalysts, reinforcing the idea that Ru acts as a promoter in the hydrogenolysis reaction [88]. In order to explore the reaction mechanism of hydrogenolysis using Ir-ReOx/SiO2 catalysts, Varghese et al. carried out an experimental and simulation study in liquid phase batch reactors. The characterization of the catalysts by photoelectron spectroscopy (XPS) revealed the presence of metallic Ir and ReOx particles with multiple oxidation states (0, +4, +7). A catalyst structural model was developed based on this study, considering that on metallic Ir, nanoparticles are dispersed ReOx particles with an average oxidation state of +4 for Re. Their results also revealed that once the catalyst was reduced, the ReOx particles constituted Brønsted acid sites with Ir-H3Re3O6 type clusters [89].
In addition to SiO2, other supports were studied to support the Ir-ReOx couplet. In this regard, Liu et al. investigated Ir-ReOx catalysts supported on anatase TiO2, rutile TiO2, C, ZrO2, CeO2, SiO2, Al2O3, and MgO. Of all the materials used as supports, rutile TiO2 was  In the presence of another metal such as Ru, the Ir-ReOx/SiO2 catalysts showed enhanced activity levels toward the formation of 1,3-PG. The presence of Ru at a given concentration dissociates H2 molecules and transfers H species via a spillover mechanism toward the Ir-ReOx particles. These species form protons and facilitate the formation of 1,3-PG ( Figure 10) [80]. Surprisingly, Ru-Re/SiO2 catalysts were found to be less active and selective than the Ir-ReOx/SiO2 catalysts, reinforcing the idea that Ru acts as a promoter in the hydrogenolysis reaction [88]. In order to explore the reaction mechanism of hydrogenolysis using Ir-ReOx/SiO2 catalysts, Varghese et al. carried out an experimental and simulation study in liquid phase batch reactors. The characterization of the catalysts by photoelectron spectroscopy (XPS) revealed the presence of metallic Ir and ReOx particles with multiple oxidation states (0, +4, +7). A catalyst structural model was developed based on this study, considering that on metallic Ir, nanoparticles are dispersed ReOx particles with an average oxidation state of +4 for Re. Their results also revealed that once the catalyst was reduced, the ReOx particles constituted Brønsted acid sites with Ir-H3Re3O6 type clusters [89].
In addition to SiO2, other supports were studied to support the Ir-ReOx couplet. In this regard, Liu et al. investigated Ir-ReOx catalysts supported on anatase TiO2, rutile TiO2, C, ZrO2, CeO2, SiO2, Al2O3, and MgO. Of all the materials used as supports, rutile TiO2 was In order to explore the reaction mechanism of hydrogenolysis using Ir-ReO x /SiO 2 catalysts, Varghese et al. carried out an experimental and simulation study in liquid phase batch reactors. The characterization of the catalysts by photoelectron spectroscopy (XPS) revealed the presence of metallic Ir and ReO x particles with multiple oxidation states (0, +4, +7). A catalyst structural model was developed based on this study, considering that on metallic Ir, nanoparticles are dispersed ReO x particles with an average oxidation state of +4 for Re. Their results also revealed that once the catalyst was reduced, the ReO x particles constituted Brønsted acid sites with Ir-H 3 Re 3 O 6 type clusters [89].
In addition to SiO 2 , other supports were studied to support the Ir-ReO x couplet. In this regard, Liu et al. investigated Ir-ReO x catalysts supported on anatase TiO 2 , rutile TiO 2 , C, ZrO 2 , CeO 2 , SiO 2 , Al 2 O 3 , and MgO. Of all the materials used as supports, rutile TiO 2 was found to be the most active. The results were attributed to the stabilization and dispersion of the Ir-ReO x particles on the rutile TiO 2 surface, which allowed for an increase in the number of active sites. In these, the Ir particles interacted with the partially oxidized ReO x clusters (average Re oxidation state: +3) covering almost the entire rutile TiO 2 surface [81].
From the zeolite group, Wan et al. prepared IrO x /H-ZSM-5 catalysts in the absence of ReO x . Their results showed that the activity is a function of Ir dispersion, Ir (0)/Ir (III) ratio, and Brønsted acid site content. The activity was found to be linear as a function of the content of Brønsted acid sites while the selectivity to 1,3-PG increased with the presence of this type of site, denoted as Ir-O(H)-H-ZSM-5. These sites are generated by the synergistic interaction between IrO x and H-ZSM-5 species through a hydrogen spillover. The authors proposed a direct mechanism for the formation of 1,3-PG on IrO x /H-ZSM-5, considering that glycerol adsorption occurs on IrO x particles and H 2 activation takes place on the Si-Al sites of the H-ZSM-5 zeolite (Figure 11) [82].
From the zeolite group, Wan et al. prepared IrOx/H-ZSM-5 catalysts in the absence of ReOx. Their results showed that the activity is a function of Ir dispersion, Ir (0)/Ir (III) ratio, and Brønsted acid site content. The activity was found to be linear as a function of the content of Brønsted acid sites while the selectivity to 1,3-PG increased with the presence of this type of site, denoted as Ir-O(H)-H-ZSM-5. These sites are generated by the synergistic interaction between IrOx and H-ZSM-5 species through a hydrogen spillover. The authors proposed a direct mechanism for the formation of 1,3-PG on IrOx/H-ZSM-5, considering that glycerol adsorption occurs on IrOx particles and H2 activation takes place on the Si-Al sites of the H-ZSM-5 zeolite (Figure 11) [82].

Recently, Chanklang et al. prepared Ir-ReOx catalysts supported on H-ZSM-5, which
showed high metal dispersion with a decrease in the Ir particle size by the addition of Re. The results indicated that the activity is a function of the Ir metal content and Re/Ir molar ratio. However, even by optimizing the reaction conditions (temperature, reaction time, and catalyst concentration), the catalysts were found to be poorly selective at 1,3-PG (~19%) with very low yields (~2.8%) due to excessive hydrogenolysis of the glycol to form 1-POH. The authors proposed a mechanism similar to the direct mechanism of hydrogenolysis used to explain the results in Ir-ReOx/SiO2 ( Figure 12) [83]. Recently, Chanklang et al. prepared Ir-ReO x catalysts supported on H-ZSM-5, which showed high metal dispersion with a decrease in the Ir particle size by the addition of Re. The results indicated that the activity is a function of the Ir metal content and Re/Ir molar ratio. However, even by optimizing the reaction conditions (temperature, reaction time, and catalyst concentration), the catalysts were found to be poorly selective at 1,3-PG (~19%) with very low yields (~2.8%) due to excessive hydrogenolysis of the glycol to form 1-POH. The authors proposed a mechanism similar to the direct mechanism of hydrogenolysis used to explain the results in Ir-ReO x /SiO 2 ( Figure 12) [83]. Other Ir-Re systems on mesoporous silica and SiO2-Al2O3 supports were also studied from the structural point of view.
Deng et al. studied the reduction of Ir-Re (Ir:Re = 1) catalysts supported on a mesoporous silica of the KIT-6 type in the range of 400 to 700 °C. Their results showed an improvement in glycerol conversion with an increasing reduction temperature up to 600 °C Other Ir-Re systems on mesoporous silica and SiO 2 -Al 2 O 3 supports were also studied from the structural point of view.
Deng et al. studied the reduction of Ir-Re (Ir:Re = 1) catalysts supported on a mesoporous silica of the KIT-6 type in the range of 400 to 700 • C. Their results showed an improvement in glycerol conversion with an increasing reduction temperature up to 600 • C due to an increase in the number of acid sites that produced a higher Ir-Re interaction. For temperatures between 600 and 700 • C, a decrease in the activity of these catalysts was observed due to a growth in the size of the metal particles [84]. In another work by these authors, they compared the Ir-Re/KIT-6 catalyst prepared by direct reduction at 500 • C of (Ir-Re/KIT-6-R) with the same precursor, but previously calcined in air at 500 • C (Ir-Re/KIT-6-CR). The results showed that the direct reduction in a H 2 atmosphere allowed for the generation of metallic Re particles that together with Ir formed a much more active Ir-Re alloy than the Ir-ReO x phase formed during calcination, because the ReO x particles cannot be easily reduced once the oxide is formed. In the presence of an Amberlyst-15 type resin, it is possible to reach the maximum yield to 1,3-PG [90].
In Ir-Re catalysts supported on silica-alumina, Ir-Re/SiO 2 -Al 2 O 3 , it has been reported that the presence of Al promotes the dispersion of Re particles and facilitates the formation of large Ir crystals, leading to low Ir-Re interaction and, consequently, low activity levels. When the Al content in the support, SiO 2 -Al 2 O 3 (-Al), is reduced, the Re particles form an alloy with the Ir particles due to the proximity between them and their interaction, which promotes higher activity levels and selectivity to 1,3-PG ( Figure 13) [85]. Despite the efforts made with these higher complexity supports, the results show that the highest yields to 1,3-PG (~38%) were obtained with Ir-ReOx/SiO2 catalysts at moderate temperature (120 °C) and pressure (8 MPa) conditions operating in liquid phase batch reactors.

Pt Catalysts
Pt catalysts were used for the production of 1,3-PG in both the liquid and vapor phases. Tables 4 and 5 summarize the main catalysts tested in the liquid phase using batch and continuous flow reactors, respectively.
Of the set of noble metals, Ru, Pd, Rh, Ir, and Pt, the latter has shown to be the most promising one for obtaining 1,3-PG, since, unlike the other metal phases, Pt acts not only as an active site for hydrogenation, but also facilitates the cleavage of the secondary C-O bond in the glycerol molecule [91].
Studies have mainly focused on the modification of traditional supports to prepare Pt catalysts, leading to the formation of 1,3-PG. The modification of supports was carried out using W and Re species, but, even so, the yields achieved were less than 70%.
The first studies were based on the modification of Pt catalysts using Re. In this sense, the Pt-Re/C catalysts were found to be more active and selective toward the formation of 1,3-PG than monometallic Pt/C and Re/C catalysts due to the formation of a Pt-Re alloy with an average metal particle size smaller than 2 nm [92]. The same Pt-Re combination Despite the efforts made with these higher complexity supports, the results show that the highest yields to 1,3-PG (~38%) were obtained with Ir-ReO x /SiO 2 catalysts at moderate temperature (120 • C) and pressure (8 MPa) conditions operating in liquid phase batch reactors.

Pt Catalysts
Pt catalysts were used for the production of 1,3-PG in both the liquid and vapor phases. Tables 4 and 5 summarize the main catalysts tested in the liquid phase using batch and continuous flow reactors, respectively.
Of the set of noble metals, Ru, Pd, Rh, Ir, and Pt, the latter has shown to be the most promising one for obtaining 1,3-PG, since, unlike the other metal phases, Pt acts not only as an active site for hydrogenation, but also facilitates the cleavage of the secondary C-O bond in the glycerol molecule [91].
Studies have mainly focused on the modification of traditional supports to prepare Pt catalysts, leading to the formation of 1,3-PG. The modification of supports was carried out using W and Re species, but, even so, the yields achieved were less than 70%.
The first studies were based on the modification of Pt catalysts using Re. In this sense, the Pt-Re/C catalysts were found to be more active and selective toward the formation of 1,3-PG than monometallic Pt/C and Re/C catalysts due to the formation of a Pt-Re alloy with an average metal particle size smaller than 2 nm [92]. The same Pt-Re combination was supported on SiO 2 , C, and SiO 2 modified with W. The study revealed that, although Pt was completely reduced in the alloy, Re was oxidized with a distribution of oxidation states, conferring a Brønsted-type acidity to the Pt-Re particles, which influenced the selectivity to 1,3-PG [93].
With respect to the use of HTPA, H 4 SiW 12 O 40 was found to be the most effective in producing 1,3-PG, either by the addition to the reaction medium in the presence of Pt/Al 2 O 3 catalysts [94], or by the modification of supports such as SiO 2 [118], ZrO 2 [119,120], Al 2 O 3 [95], and ordered mesoporous Al 2 O 3 (m-Al 2 O 3 ) [96]. The results indicated that the modification of these supports with the heteropolyacid enhanced the dispersion of Pt and conferred the appropriate acidity for the formation of the intermediate in the production of 1,3-PG. In this respect, it has been verified that there is a linear correlation between the yield to 1,3-PG and the surface concentration of Brønsted acid sites provided by the heteropolyacids. This trend was verified using Pt-H 4 SiW 12 O 40 /ZrO 2 [119], Pt-H 3 PW 12 O 40 -ZrO 2 [120], and Pt-H 3 PMo 12 O 40 /ZrO 2 [120].
With respect to modification with tungsten oxides (WO x ), it has been shown that the presence of these oxides endows the support with Brønsted acid sites for the adsorption of glycerol and the subsequent formation of an alkoxide intermediate in the formation of 1,3-PG, following the direct mechanism of hydrogenolysis [97]. The tungsten redox cycle, W +6 ↔ W +5 , further allows in situ regeneration of the surface sites after the formation of 1,3-PG [129]. In the Pt/WO x catalysts, the interaction between Pt particles and metallic H 2 produces a hydrogen spillover from Pt to WO x . The presence of oxygen vacancies in the WO x facilitates the mechanism of heterolytic cleavage of H 2 into H + and H − ions, which are necessary in hydrogenolysis for the formation of 1,3-PG from the alkoxide intermediate ( Figure 14) [63].
In situ ATR-IR studies revealed that WO x plays three fundamental roles in the formation of 1,3-PG: (1) adsorption of the primary -OH group of glycerol; (2) provision of H+; and (3) stabilization of the intermediate carbocation in the formation of 1,3-PG [66].
Recently, Yang et al. carried out characterization studies by photoelectron spectroscopy and hydrogen chemisorption and determined that the catalytic activity of Pt/WO x was from the oxygen vacancies found in the WO x species. By preparing Pt/Al-WO x catalysts, they determined that by modulating the content of atomically dispersed AlO x species on the WO x support, it was possible to generate a higher content of oxygen vacancies and thus promote the formation of 1,3-PG ( Figure 15) [98]. produces a hydrogen spillover from Pt to WOx. The presence of oxygen vacancie WOx facilitates the mechanism of heterolytic cleavage of H2 into H + and Hions, wh necessary in hydrogenolysis for the formation of 1,3-PG from the alkoxide interm ( Figure 14) [63].  In order to promote the activity of the Pt/WOx catalysts, studies on the addition promoters such as La [99] and Au [100] have been reported. Both strategies improved t activity and selectivity toward 1,3-PG due to the extra generation of Brønsted acid sit In the case of Pt-Au/WOx, the presence of Au enhanced the dissociative adsorption of [100]. Yang et al. characterized Pt/Au/WO3 catalysts prepared by a surface-modified de osition method and determined that the presence of Au promotes the reduction of Pt a W at low temperatures while making the dispersion of Pt uniform on the WO3 suppo These changes in the electronic properties allow for an improvement in the activity an selectivity of the catalyst [101].
With respect to the Pt/WOx/ZrO2 catalysts, their activity in the isomerization of kanes is well-known [130][131][132][133]. On the surface of these catalysts, H2 molecules are tran formed into protons (H + ) and hydrides (H -) via heterolytic cleavage [134,135]. The hydr genolysis reaction of glycerol to form 1,3-PG consists of a substitution reaction, in whi an H atom replaces the secondary -OH group in the glycerol molecule. For this reason possible sequence is the two-step reaction of glycerol with H + and H -, which produc water and the substitution of the secondary -OH group by H [136,137]. Following th In order to promote the activity of the Pt/WO x catalysts, studies on the addition of promoters such as La [99] and Au [100] have been reported. Both strategies improved the activity and selectivity toward 1,3-PG due to the extra generation of Brønsted acid sites. In the case of Pt-Au/WO x , the presence of Au enhanced the dissociative adsorption of H 2 [100]. Yang et al. characterized Pt/Au/WO 3 catalysts prepared by a surface-modified deposition method and determined that the presence of Au promotes the reduction of Pt and W at low temperatures while making the dispersion of Pt uniform on the WO 3 support. These changes in the electronic properties allow for an improvement in the activity and selectivity of the catalyst [101].
With respect to the Pt/WO x /ZrO 2 catalysts, their activity in the isomerization of alkanes is well-known [130][131][132][133]. On the surface of these catalysts, H 2 molecules are transformed into protons (H + ) and hydrides (H − ) via heterolytic cleavage [134,135]. The hydrogenolysis reaction of glycerol to form 1,3-PG consists of a substitution reaction, in which an H atom replaces the secondary -OH group in the glycerol molecule. For this reason, a possible sequence is the two-step reaction of glycerol with H + and H − , which produces water and the substitution of the secondary -OH group by H [136,137]. Following this idea, Pt/WO x /ZrO 2 catalysts [62,91,102] were used for the hydrogenolysis of glycerol to 1,3-PG.
Compared to other traditional supports, ZrO 2 is one of the most interesting to combine with Pt as the metal phase and to generate Brønsted acid sites. For example, the modification of ZrO 2 with sulfonic groups allows for the preparation of Pt catalysts that showed yields of 55% to 1,3-PG due to the generation of Brønsted acid sites [110]. It has also been indicated that the tetragonal phase of ZrO 2 is more active than the monoclinic phase to be used as a support, since it generates higher metal dispersion and Brønsted acid sites of higher acidic strength [102].
In comparison to catalysts such as Pt/WO 3 /TiO 2 , Pt/WO 3 /HY, and Pt/WO 3 / AlMCM-41, the Pt/WO 3 /ZrO 2 catalyst proved to be the best alternative, reaching the maximum yield to 1,3-PG [91]. In these catalysts, the formation of 1,3-PG is structurally sensitive to the degree of polymerization of WO x species. The Pt-(WO x ) n -H type active sites provide a strong Brønsted acidity that promotes the formation of 1,3-PG [111].
Pt/WO x /SiO 2 /ZrO 2 catalysts resulted in being more active and selective to 1,3-PG due to a promoting effect of SiO 2 that favors not only the dispersion of Pt, but also of the WO x species [124]. This effect has been observed in similar Pt/WO x /TiO 2 /SiO 2 catalytic systems [97]. The presence of an adequate SiO 2 content favors the transformation of crystalline WO 3 species into polytungstate species, which are more active and allow for an increase in the yield to 1,3-PG. When the SiO 2 content exceeds the optimum, the polytungstate species are transformed into monotungstate species that decrease the yield to 1,3-PG (Figure 16a). This leads to obtaining a maximum yield to 1,3-PG as a function of the SiO 2 content (Figure 16b) [124].
Zhao et al. proved that the presence of bibridged (OH II ) and tribridged hydroxyl groups (OH III ) on the SiO 2 /ZrO 2 support could react with WO x species forming Zr-O-W bands and retarding the agglomeration of WO x species. The high dispersion of WO x species in conjunction with Zr-O-W bonds benefit the formation of Brønsted acid sites that promote the formation of 1,3-PG [112].
Xi et al. prepared a series of Pt/WO x -ZrO 2 -ZrO 2 -TiO 2 catalysts from an evaporative assembly method using different Ti/Zr molar ratios. The authors detected the presence of ZrTiO 4 and Ti 2 ZrO 6 phases, responsible for the activity in the hydrogenolysis of glycerol. For a Ti/Zr = 7:3 molar ratio, the catalyst showed the maximum yield to 1,3-PG (~25.3%) [125].
The modification of Pt/WO x /ZrO 2 catalysts with Li 2 B 4 O 7 in the 0.5-2 wt.% range allowed for a better dispersion of the Pt particles and increased the total number of acid sites, thus increasing the interaction between the metal phase and the support. At 150 • C and 4 MPa H 2 , the catalyst with 2 wt.% Li 2 B 4 O 7 allowed for obtaining the maximum yield to 1,3-PG (~42.7%), being stable for more than 200 h [126].
Doping Mn into the Pt/WO x /ZrO 2 catalysts allowed for the generation of a large number of WO x species with some intermediate degree of polymerization, which increased the number of Pt-(WO x ) n -H sites promoting the formation of 1,3-PG [111]. The same effect was found with the incorporation of Mg to the Pt/WO x /ZrO 2 catalysts. In this case, the formation of small WO x species with some degree of polymerization was active for the formation of 1,3-PG [127].
The preparation of Pt/WO 3 /ZrP catalysts also led to good yields to 1,3-PG, but lower in comparison to Pt/WO x/ ZrO 2 catalysts. The co-presence of Pt and W in the catalyst composition led to a synergistic enhancement in the catalytic activity and 1,3-PDO selectivity. Highly dispersed Pt nanoparticles in contact with polytungstate-like WO x species were active centers for the hydrogenolysis reaction [113]. [97]. The presence of an adequate SiO2 content favors the transformation of crystalline WO3 species into polytungstate species, which are more active and allow for an increase in the yield to 1,3-PG. When the SiO2 content exceeds the optimum, the polytungstate species are transformed into monotungstate species that decrease the yield to 1,3-PG (Figure 16a). This leads to obtaining a maximum yield to 1,3-PG as a function of the SiO2 content (Figure 16b) [124].  With respect to the modification of Al 2 O 3 with WO x , it has been reported that the surface density of W determines the type of W species on the support surface and thus the selectivity to 1,3-PG. In the Pt/WO x /Al 2 O 3 catalysts, the presence of monotungstate, polytungstate, and mesoporous WO 3 species generates the Brønsted acid sites necessary for the cleavage of the secondary C-O bond of the glycerol molecule [121]. In this regard, it has been reported that the transformation of WO 3 to polytungstate is required to achieve the highest yields to 1,3-PG [124]. The combination of polytungstate species with highly dispersed Pt particles leads to the formation of alkoxide intermediates and their subsequent hydrogenation to form 1,3-PG ( Figure 17) [65]. selectivity to 1,3-PG. In the Pt/WOx/Al2O3 catalysts, the presence of monotungstate, polytungstate, and mesoporous WO3 species generates the Brønsted acid sites necessary for the cleavage of the secondary C-O bond of the glycerol molecule [121]. In this regard, it has been reported that the transformation of WO3 to polytungstate is required to achieve the highest yields to 1,3-PG [124]. The combination of polytungstate species with highly dispersed Pt particles leads to the formation of alkoxide intermediates and their subsequent hydrogenation to form 1,3-PG ( Figure 17) [65]. It has also been reported that an excessive number of tungsten species can lead to the formation of undesired products such as 1-POH, 2-POH, MeOH, and gases [138]. Thus, a proper balance between Pt, W species, and support is required, leading to the need for an optimization of the W content in the catalyst. In this regard, Lei et al. studied various It has also been reported that an excessive number of tungsten species can lead to the formation of undesired products such as 1-POH, 2-POH, MeOH, and gases [138]. Thus, a proper balance between Pt, W species, and support is required, leading to the need for an optimization of the W content in the catalyst. In this regard, Lei et al. studied various compositions of Pt and WO x species on Al 2 O 3 . Their results determined that for a fixed Pt content (2 wt.%), an increase in the W concentration up to 7.5 wt.% causes an increase in the selectivity to 1,3-PG; if the W concentration continues to increase, the selectivity to 1,3-PG decreases by the generation of products such as 1-POH. Similarly, maintaining a fixed W content (12.9 wt.%), an increase in the Pt concentration up to 6 wt.% causes an increase in glycerol conversion, while for higher contents of the metal, there is a drop in activity due to a decrease in the metal dispersion. The results of this work indicate that for each W content and optimum Pt content (or vice versa), it is required to achieve the highest levels of glycerol conversion and selectivity to 1,3-PG, which implies maximizing the yield to the glycol of interest, as shown in Figure 18 [128]. tions 2022, 3, FOR PEER REVIEW 29 compositions of Pt and WOx species on Al2O3. Their results determined that for a fixed Pt content (2 wt.%), an increase in the W concentration up to 7.5 wt.% causes an increase in the selectivity to 1,3-PG; if the W concentration continues to increase, the selectivity to 1,3-PG decreases by the generation of products such as 1-POH. Similarly, maintaining a fixed W content (12.9 wt.%), an increase in the Pt concentration up to 6 wt.% causes an increase in glycerol conversion, while for higher contents of the metal, there is a drop in activity due to a decrease in the metal dispersion. The results of this work indicate that for each W content and optimum Pt content (or vice versa), it is required to achieve the highest levels of glycerol conversion and selectivity to 1,3-PG, which implies maximizing the yield to the glycol of interest, as shown in Figure 18 [128]. Zhao et al. developed a strategy to prepare Pt catalysts supported on Al2O3 containing highly dispersed WOx species by employing high-temperature heat treatment. In their work, these authors attributed the catalytic efficiency to the strong interaction between isolated WO4 species and the Pt metal phase, which would be responsible for generating Zhao et al. developed a strategy to prepare Pt catalysts supported on Al 2 O 3 containing highly dispersed WO x species by employing high-temperature heat treatment. In their work, these authors attributed the catalytic efficiency to the strong interaction between isolated WO 4 species and the Pt metal phase, which would be responsible for generating Brønsted acid sites in situ, according to the results of the IR spectroscopy and NH 3 adsorption-desorption [106].
In order to improve the properties of the Pt/WO x /Al 2 O 3 catalyst, Wen et al. studied the modification of the metal phase with Ru. The authors found that the presence of Ru favored the redox cycle W +6 ↔ W +5 , which allowed for an improvement in the adsorption of H 2 and an increase in the number of Brønsted acid sites of medium acidic strength, leading to the attainment of a higher yield to 1,3-PG [114].
Similar results were obtained with the Pt-Au/WO x /Al 2 O 3 catalysts. In these systems, the addition of Au, which exists on the catalytic surface mostly as single atoms, enhances the hydrogen spillover capacity, decreasing the interaction between the Pt and WO x species, promoting Pt dispersion (~2 nm) and increasing its exposed area. Since the spillover capacity is directly linked to the number of Brønsted acid sites, the presence of Au favors the production of 1,3-PG [115].
For catalysts of similar characteristics, but based on bohemite (AlOOH), a reaction mechanism through alkoxide type intermediate species has been proposed, in whose formation Pt, WO x , and AlOOH type oxides are cooperatively involved. Initially, glycerol is adsorbed on the surface sites of the bohemite through the primary -OH group, forming alkoxide species of the HOCH 2 CH(OH)CH 2 OAlO x type. Next, the secondary -OH group of the alkoxide species is protonated on the acid sites of the WO x and dehydrates, producing a secondary carbocation. This carbocation is attacked by hydride species generated on the WO x sites due to the H 2 spillover produced by the interaction between the Pt particles and the metallic H 2 . Finally, the alkoxide species are hydrolyzed to generate 1,3-PG and the catalytic cycle is completed. On the other hand, the surface of the bohemite plays a fundamental role as it stabilizes the Pt particles and the W oxides [64]. With this catalyst, it is possible to reach a 66% yield to 1,3-PG, which is the highest yield reported to date.
Recently, yields very close to that value (~65%) have been reported by Cheng et al., who prepared Pt catalysts (Pt/W-MCFs) supported on tungsten-doped siliceous mesocellular foams (W-MCFs). The synergy between the Pt nanoparticles and the tetrahedral sites of W V O 4 species at the oxide-metal interface were responsible for the catalyst activity as well as the 3-dimensional continuous large pore structure of the W-MCFs, which facilitates the entry of the reactant and the escape of the product [116].
With respect to the Pt/WO x /SBA-15 catalysts, the addition of W produces WO 4 , which is incorporated into the SBA-15 matrix as a Lewis acid site. The H 2 adsorbed on Pt is transferred by spillover to the W oxide, which is transformed into a Brønsted acid site, allowing for the formation of the intermediate and then of 1,3-PG [67].
Feng et. al. also prepared Pt/SBA-15 catalysts, but modified with W and Al, which were homogeneously incorporated into the SBA-15 matrix using a one-step synthesis method under an acidic medium. The joint presence of W and Al allowed for an increase in the synergism effect between the Lewis and Brønsted acid sites, which were tightly related to the incorporation of W and Al species on the modified SBA-15. According to these authors, glycerol adsorbs on the catalytic surface at the WO x sites to form a terminal alkoxide, which then dehydrates and hydrogenates to finally form 1,3-PG. Naturally, the adsorption and activation of H 2 takes place on the Pt metal sites (Figure 19) [107].
Recently, Zhou et al. tested Pt/WO x /TiO 2 catalysts that were found to be active and selective to 1,3-PG, showing 37% yields for 57% glycerol conversions at 120 • C and 4 MPa H 2 after 8 h of reaction. Modification of these catalysts with sulfate ions allowed them to achieve full glycerol conversion while maintaining the overall yield to 1,3-PG under the same reaction conditions. The result was attributed to the increase in Brønsted acid sites as well as a better dispersion of Pt particles and WO x species [117].
were homogeneously incorporated into the SBA-15 matrix using a one-step synthesis method under an acidic medium. The joint presence of W and Al allowed for an increase in the synergism effect between the Lewis and Brønsted acid sites, which were tightly related to the incorporation of W and Al species on the modified SBA-15. According to these authors, glycerol adsorbs on the catalytic surface at the WOx sites to form a terminal alkoxide, which then dehydrates and hydrogenates to finally form 1,3-PG. Naturally, the adsorption and activation of H2 takes place on the Pt metal sites (Figure 19) [107]. Recently, Zhou et al. tested Pt/WOx/TiO2 catalysts that were found to be active and selective to 1,3-PG, showing 37% yields for 57% glycerol conversions at 120 °C and 4 MPa H2 after 8 h of reaction. Modification of these catalysts with sulfate ions allowed them to achieve full glycerol conversion while maintaining the overall yield to 1,3-PG under the same reaction conditions. The result was attributed to the increase in Brønsted acid sites as well as a better dispersion of Pt particles and WOx species [117].
Other more current systems based on Pt-WOx have been supported on Ta2O5. The calcination of these catalysts at high temperature (900 °C) has allowed for high dispersions of tungsten oxides to be obtained with a low degree of polymerization, increasing the oxygen vacancies on their surface. This generates a higher dispersion of the Pt particles, maximizing the Pt-O-W interfacial bonding, and leads to obtaining higher yields to 1,3-PG [108,109].
With respect to obtaining 1,3-PG in the vapor phase, Table 6 summarizes the main Pt catalysts employed in continuous flow reactors. Other more current systems based on Pt-WO x have been supported on Ta 2 O 5 . The calcination of these catalysts at high temperature (900 • C) has allowed for high dispersions of tungsten oxides to be obtained with a low degree of polymerization, increasing the oxygen vacancies on their surface. This generates a higher dispersion of the Pt particles, maximizing the Pt-O-W interfacial bonding, and leads to obtaining higher yields to 1,3-PG [108,109].
The Pt/WO 3 /Al 2 O 3 catalysts that had demonstrated yields to 1,3-PG close to 40% in the liquid phase showed low yields to 1,3-PG in the vapor phase (14%), probably due to the thermal level in combination with the surface acidity, which promotes the formation of undesired products. In this regard, the activity results indicated that the reaction proceeds via an initial dehydration of glycerol to form acrolein, which is then rehydrated to form 3-HPA, an intermediate of 1,3-PG [139].
Surprisingly, the modification of SBA-15 with WO 3 allowed for the preparation of Pt catalysts that showed higher yields to 1,3-PG than Pt/WO 3 /Al2O 3 (36%), probably due to lower acidity and less severe reaction conditions. In these systems, the activity of the Brønsted acid sites present in WO 3 is combined with the hydrogenation activity of Pt [140]. Similar yields were obtained with Pt/AlPO 4 catalysts due to the presence of weak Brønsted acid sites in AlPO 4 and the dispersion of Pt particles [141].
Other Pt [142] and Pt-Cu [143] based catalysts on protonated mordenite supports (Pt/H-MOR, Pt-Cu/H-MOR) showed superior performance for the formation of 1,3-PG with yields of 46-52% at 0.1 MPa H 2 and 1 h −1 (WHSV) in the 225-215 • C range, respectively. In these systems, the dispersion of Pt combined with the concentration of Brønsted acid sites of the support promote the selective formation of 1,3-PG [142]. In the case of the bimetallic catalyst, the presence of Cu decreases the C-C cleavage side reactions promoted by Pt [143]. The highest yields to 1,3-PG were obtained with a Pt catalyst on sulfuric acid activated montmorillonite (Pt/S-MMT) (~58%) at 200 • C and 0.1 MPa H 2 [144].

Engineering Aspects and Effect of Reactor Parameters
The glycerol hydrogenolysis reaction to 1,3-PG can be carried out in either the liquid or vapor phase. The liquid phase condition requires moderate temperatures (120-240 • C) and high pressures (2-10 MPa) of H 2 or N 2 depending on the case. Generally, a H 2 atmosphere is used due to the requirement in the hydrogenation stage, although there are some studies where N 2 is used and H 2 donor substances are included [145][146][147]. The liquid phase condition involves higher fixed costs associated with the reaction equipment due to the high pressures used, but has the advantage of not requiring vaporization of the reaction mixture [47].
Batch reactors and continuous flow reactors can be used to carry out the hydrogenolysis reaction in the liquid phase. The use of batch reactors allows for experimental data to be obtained for the future formulation of kinetic models, the study of temperature, pressure, reaction time, and the concentration of reagents and the catalyst being of relevance as operating variables. One of the disadvantages of batch reactors is the increase in the concentration of secondary products as the reaction time increases, which affects the yield to glycols [148]. Recently, Jean et al. recommended the use of continuous flow trickle bed reactors to carry out liquid phase hydrogenolysis [149]. This type of reactor offers several advantages, mainly due to the low contact times between the glycerol and the active sites of the catalyst [150][151][152], and thus the catalysts show higher stability.
On the other hand, the vapor phase condition requires low pressures (0.1-0.5 MPa) and higher temperatures (200-260 • C) and fixed-bed continuous flow reactors have been used to carry this out. Disadvantages include the need to vaporize the fed glycerol solution, which requires high temperatures and high energy consumption. Since the crude glycerol is in a concentration ranging from 50 to 80 wt.%, the temperatures required for vaporization could lead to the decomposition of the glycerol. Because lower H 2 pressures are used, higher selectivity values to dehydration intermediates are obtained and polymerization reactions can also be favored [153]. In these reactors, the study of temperature, pressure, space velocity, and H 2 flow rate are the main operating variables.
A detailed study is presented below of the different operating variables studied in batch reactors and in continuous flow reactors in both the liquid and vapor phase using different catalytic systems.

Effect of the Initial Glycerol Concentration
Since glycerol from biodiesel synthesis is found in aqueous solutions in the range of 50-80 wt.%, it is of relevance to study its effect on catalyst performance. Once the catalyst mass is fixed, the effect of the glycerol content has been reported for both the liquid phase condition (Table 7) and the vapor phase condition (Table 8). In general, the results indicate that increasing the glycerol concentration produces a decrease in conversion, regardless of the catalytic system considered and the operating condition of the liquid phase [55] or vapor phase [140,141]. This effect is due to the lower availability of active sites at high glycerol concentrations, since the catalyst mass does not change but the glycerol content does. On the other hand, selectivity to 1,3-PG depends on the glycerol concentration and decreases with an increasing concentration in the presence of Ru [55] and Pt [141,142] based catalysts. SV-space velocity; Y 1,3-PG -yield to 1,3-PG.

Effect of Initial Water Concentration
Due to the presence of water in crude glycerol (20-50 wt.%), its effect has also been studied for different catalytic systems.
Regarding the initial water content in the glycerol solution, there are several aspects to consider. It has been reported that a minimum water content in the starting glycerol solution is necessary, since its presence inhibits side reactions of glycerol dehydration to undesired products [154] such as the dehydration of 1,3-PG to allyl alcohol [155] as well as condensation reactions between the glycerol and side products such as EG, EtOH, and MeOH [156]. Moreover, in continuous flow steam-phase reactors, it is desirable to minimize the water content to produce energy savings in vaporization [155], but its presence allows for vaporization of the glycerol solution to be achieved at lower temperature.
Several studies have reported that when the water content is very high, the hydrogenolysis of glycerol can lead to the formation of C-C bond cleavage products such as EtOH [154]. From another point of view, a techno-economic study has been reported to carry out the purification of 1,2-PG by distillation in stages. The first stage consists in separating the water from the liquid products, then separating the unreacted glycerol and finally the 1,2-PG from the rest of the products, especially EG. The economic study determined that the separation of water in the first distillation stage was the costliest step of the whole purification [157].
Regarding the performance of the catalysts, since water is generated as a by-product during hydrogenolysis, it is better to work with low water contents in the starting solution to shift the equilibrium toward product formation.
In this regard, Nakagawa et al. studied the effect of increasing the glycerol concentration from 5 to 80 wt.% by decreasing the water content from 95 to 20 wt.% using Ir-ReO x /SiO 2 catalysts in the presence of H-ZSM-5. Their results showed an increase in glycerol conversion from 31% to 51% with a slight increase in selectivity to 1,3-PG [78]. The same effect was reported by Varghese et al. who, employing the Ir-ReO x /SiO 2 catalyst, observed an increase in glycerol conversion from 40 to 53% with a maintenance in selectivity at 1,3-PG (~40%) when the water content decreased from 98 to 80 wt.% at 120 • C, 8 MPa, and 24 h reaction [89].

Effect of the Solvent
Since water has a negative effect on the activity of the catalysts and their selectivity to 1,3-PG, the replacement of water by other solvents has been a focus of study. Among the solvents used in the hydrogenolysis reaction, aprotic and aprotic solvents have been used.
The use of protic solvents, on the other hand, has the advantage that many of them are H 2 donor substances that allow H 2 to be generated in situ and thus improve the performance of the catalysts. Moreover, it has been reported that in the presence of protic solvents such as MeOH and EtOH, the H 2 solubility is much higher than in water [158]. There are, however, some drawbacks. Since glycerol and hydrogen donor substances compete for the same active sites in the hydrogenolysis process, there should be a balance between the amount of glycerol and the amount of hydrogen donor substance to achieve high conversion and selectivity to the desired product [146]. In addition, the use of protic solvents such as EtOH and MeOH favors the formation of condensation products such as glyceryl ethers [152].
For the production of 1,2-PG, MeOH, EtOH, 2-POH, formic acid [145][146][147]157,159,160], and EG [161] have been used as protic solvents. Strikingly, in the production of 1,3-PG, only the use of MeOH [152], water, and EtOH [97,152,162,163] have been evaluated. Of all of them, water and EtOH have been shown to be the best solvents due to their ability to ionize and generate H + , which favor the formation of, 1,3-PG, since they act as Brønsted acid sites on the surface of the catalysts, as it is in the case of Pt/WO 3 -ZrO 2 . These sites allow for the anchoring of the central -OH group on the glycerol molecule and its subsequent cleavage to produce 1,3-PG. The generation of H + on the catalytic surface avoids steric hindrance of the anchoring of the central -OH group on the glycerol molecule unlike other acidic sites, which benefits the formation of 1,3-PG [162,163]. In this regard, Gong et al. reported improved glycerol conversion and selectivity to 1,3-PG in the presence of water and EtOH as solvents, employing a Pt/WO 3 /TiO 2 /SiO 2 catalyst at 180 • C and 5.5 MPa H 2 [97]. Cai et al. reported similar results employing a Co-Al catalyst, obtaining the best selectivity to 1,3-PG in the presence of water due to the H + transfer capacity of water and its high polarity [152].

Effect of Temperature
The kinetics of the hydrogenolysis reaction is a function of temperature and, therefore, the formation of 1,3-PG, 1,2-PG and other products such as MeOH, EtOH, and EG depends on this operating variable.
It has been found that an increase in temperature causes an increase in glycerol conversion with a loss in selectivity to 1,3-PG due to the formation of side products such as 1-POH, EG, EtOH, MeOH, and gases.
It is worth mentioning that temperature is also a critical parameter in reactor operation, and severe thermal conditions can lead to high energy demands, which increases the operating costs of the hydrogenolysis process [164]. Tables 9 and 10 show the thermal effect study for different catalysts. For each of them, the tables also show the thermal range studied, the operating conditions under which the study was carried out, and the optimum temperature that maximizes the yield to 1,3-PG.  SV-space velocity; Y 1,3-PG -yield to 1,3-PG.

Effect of Pressure
The effect of pressure was studied using batch reactors (Table 11) and continuous flow reactors (Table 12).  In the liquid phase, using batch reactors, it has been reported that an increase in H 2 pressure generates an increase in glycerol conversion as well as an increase in selectivity to 1,3-PG, using catalysts based on noble metals such as Pt [66,97,101,107,114,117,166,167] and Ir [60,82,83] and non-noble metals such as Ni [165].
The effect of pressure on the conversion and selectivity was also tested in the liquid phase using continuous flow reactors of the trickle bed type and the results obtained were similar to those reported in batch reactors [62]. In general, the results obtained have been attributed to a higher solubility of H 2 in the liquid phase, which allows for hydrogenation of the intermediate in the formation of 1,3-PG.
Increasing the pressure generates two benefits. On one hand, it allows the liquid phase condition to be maintained, and on the other hand, it increases the H 2 solubility, improving the selectivity to 1,3-PG. However, reactors operating at higher pressures presuppose not only a higher fixed cost associated with the equipment, but also a higher operating cost and potential safety risks during H 2 handling [47,164].
Very few articles have been published in which the effect of pressure has been studied when the reaction is conducted in the vapor phase, and they report an increase in the conversion and selectivity at 1,3-PG [118,155] due to the effect of pressure. Table 13 shows the results of the H 2 flux study in both the liquid phase (trickle bed reactors) and vapor phase (fixed bed reactors) conditions by using different catalytic systems. The results have indicated that an increase in H 2 flux generates an increase in glycerol conversion due to increased H 2 availability at the catalytic surface, which generates the increased hydrogenation of reaction intermediates to form 1,3-PG. It has been observed that in some cases, the selectivity to 1,3-PG increases with increasing H 2 flux up to a certain point, after which it decreases due to the generation of side products [140,141,143].

Effect of Reaction Time
The effect of reaction time has been studied by several authors using batch reactors operating in the liquid phase condition (Table 14). Naturally, the conversion of glycerol in batch reactors increases as the reaction time increases. This effect was reported for several catalytic systems based on noble metals such as Pt [97,107,109,114,116,166,167] and Ir [59,80,82,83,89], and on non-noble metals such as Ni [165] and Co [77].
As the conversion increases, there is a decrease in selectivity to 1,3-PG due to the generation of side products such as 1-POH and propane. Since the conversion increases as the reaction time increases, the selectivity at 1,3-PG decreases, and there is an optimum time to obtain the maximum yield to these products.

Effect of Space Velocity
For continuous flow reactors, both in the liquid phase [118] and in the vapor phase condition [139,140,142,144,155], the special velocity (WHSV) has been studied using different catalytic systems (Table 15). It has been reported that high WHSV values lead to lower conversions due to low contact times [118,139,140,142,144]. In some cases, high WHSV values have led to obtaining high selectivity to 1,3-PG [118], while in other cases, the selectivity values have been low [140,[142][143][144]155] due to the formation of side products due to high contact times. In a number of studies, it has been indicated that there is an optimum WHSV value for which the yield to glycols is the maximum [118,155].

Stability and Reusability of Catalysts
From previous sections, it can be concluded that Ir and Pt based catalysts are promising for the formation of 1,3-PG, in particular, those consisting of Ir-ReO x /SiO 2 and Pt/WO x supported on ZrO 2 , SBA-15, and Al 2 O 3 .
In this section, we discuss the stability of such catalytic systems and the possibility of their reuse in successive reaction cycles, given their potential use for an industrial application.

Ir-ReO x /SiO 2 Catalysts
Luo et al. reported the stability of Ir-ReO x /SiO 2 catalysts in liquid-phase continuous flow reactors employed in the form of egg-shell pellets. The catalytic thickness in these systems was controlled using an ethanol solution with different concentrations (20-30 wt.%) to perform the impregnation, first with the H 2 IrCl 6 precursor and then with the NH 4 ReO 4 precursor.
The catalyst showed a conversion of 60% with a selectivity to 1,3-PG of 31% at 130 • C and 8 MPa H 2 . After 36 h of reaction, the glycerol conversion reached a maximum (~65%) and then progressively decreased to reach half that value only at 468 h of reaction. In the first 36 h of reaction, the selectivity to 1,3-PG increased from 31 to 40% and after 500 h of reaction, the yield to 1,3-PG only decreased from 19 to 13%. The cause of deactivation in these catalysts was the agglomeration and sintering of the Ir particles, together with a loss in the interfacial area of the sites and their acidity [127]. This phenomenon was attributed to the loss of W by leaching, which is favored in the presence of water. By modifying the catalyst with Mg (0.5 wt.%), the authors avoided the deactivation phenomenon, making the catalyst stable for 400 h under the same reaction conditions.

Pt/WO x /ZrO 2 Catalysts
The presence of Mg decreased the degree of polymerization of the WO x species and also increased the stability of the W species on the ZrO 2 surface, avoiding the loss of W by leaching and therefore obtaining a more stable catalyst.
Moreover, among their results, Wang et al. proved that monoclinic ZrO 2 was more stable as a support than tetragonal ZrO 2 , since the latter tends to form monoclinic ZrO 2 under the hydrothermal conditions of the reaction [127].
In order to improve the stability, Zhu et al. added Li 2 B 4 O 7 to the Pt/WO x /ZrO 2 catalyst in the 0.5-2 wt.% range, which favored the dispersion of the Pt particles and increased the total number of acid sites, increasing the interaction between the metal phase and the support. At 150 • C and 4 MPa H 2 , the catalyst with 1 wt.% Li 2 B 4 O 7 (Pt-1LiB/WZr) allowed for the most stable catalyst with an average yield to 1,3-PG of 45% being stable for more than 200 h to be obtained [126].
The authors further studied the effect of the WHSV on the catalytic stability over the Pt-1LiB/WZr catalyst. The glycerol conversion was relatively high and stable when the WHSV was equal to 0.1 h −1 and 0.2 h −1 . However, as the WHSV increased to 0.3 h −1 , glycerol conversion decreased significantly with time evolution. The higher stability was attributed to the fact that Li 2 B 4 O 7 could restrain the reduction of tetragonal zirconium after hydrothermal treatment and reaction by strengthening the interaction between the active components and support.
Another Li modifier, based on a partial replacement of a Keggin acid (H 2 Li 2 SiW 12 O 40 ), was also used to improve the stability of Pt/WO x /ZrO 2 catalysts. At 180 • C and 5 MPa H 2 using 10 wt.% glycerol solutions, this catalyst was stable for 120 h of reaction in a continuous flow reactor operating in the liquid phase. The characterization revealed a preservation of the Keggin structure as well as of the Pt particles [119].

Pt/WO x /SBA-15 Catalysts
The Pt/WO x /SBA-15 catalysts were tested in batch reactors operating in the liquid phase and were relatively stable at 150 • C and 4 MPa H 2 after four reaction cycles of 30 h in the presence of 30 wt.% glycerol solutions. The results showed a slight decrease in conversion (~3%), which was attributed to the loss of catalytic material between one reaction cycle and the next. In all cases, the selectivity to 1,3-PG remained invariant (~70%) [67].
In the vapor phase condition at 210 • C and 0.1 MPa H 2 , employing continuous flow reactors, Priya et al. also observed a slight drop in conversion (~5%), but with a decrease in selectivity to 1,3-PG being more pronounced (~8%) (Figure 20). In this case, the authors reported the formation of carbonaceous deposits on the catalytic surface, probably due to the higher severity of the operating condition [140].

Pt/WOx/Al2O3 Catalysts
Lei et al. studied the stability of Pt/WOx/Al2O3 catalysts in a liquid phase continuous flow reactor at 180 °C and 5 MPa H2 using 50 wt.% glycerol solutions. Their results initially showed a glycerol conversion of 57% with a selectivity to 1,3-PG of 40%. After 100 h of reaction, the conversion decreased to ~45% and began to decrease steadily to 25% conversion at 600 h of reaction, after which time it remained constant. The selectivity to 1,3-PG, on the other hand, remained unchanged for 700 h of reaction. The characterization results revealed a preservation in the structure of the WOx/Al2O3 support but an agglomeration of the Pt particles, leading to an increase in particle size from 1.9 to 3.1 nm. This increase in particle size led to a decrease in the Brønsted acid sites created at the Pt-WOx interface, causing catalyst deactivation [168].
Numpilai et al. carried out tests with Pt/WOx/Al2O3 catalysts in batch reactors operating in the liquid phase at 220 °C and 6 MPa H2 using 30 wt.% glycerol solutions and 5 h reaction time (Figure 21).

Pt/WO x /Al 2 O 3 Catalysts
Lei et al. studied the stability of Pt/WO x /Al 2 O 3 catalysts in a liquid phase continuous flow reactor at 180 • C and 5 MPa H 2 using 50 wt.% glycerol solutions. Their results initially showed a glycerol conversion of 57% with a selectivity to 1,3-PG of 40%. After 100 h of reaction, the conversion decreased to~45% and began to decrease steadily to 25% conversion at 600 h of reaction, after which time it remained constant. The selectivity to 1,3-PG, on the other hand, remained unchanged for 700 h of reaction. The characterization results revealed a preservation in the structure of the WO x /Al 2 O 3 support but an agglomeration of the Pt particles, leading to an increase in particle size from 1.9 to 3.1 nm. This increase in particle size led to a decrease in the Brønsted acid sites created at the Pt-WO x interface, causing catalyst deactivation [168].
Numpilai et al. carried out tests with Pt/WO x /Al 2 O 3 catalysts in batch reactors operating in the liquid phase at 220 • C and 6 MPa H 2 using 30 wt.% glycerol solutions and 5 h reaction time (Figure 21). revealed a preservation in the structure of the WOx/Al2O3 support but an agglomeration of the Pt particles, leading to an increase in particle size from 1.9 to 3.1 nm. This increase in particle size led to a decrease in the Brønsted acid sites created at the Pt-WOx interface, causing catalyst deactivation [168].
Numpilai et al. carried out tests with Pt/WOx/Al2O3 catalysts in batch reactors operating in the liquid phase at 220 °C and 6 MPa H2 using 30 wt.% glycerol solutions and 5 h reaction time (Figure 21). Figure 21. The glycerol conversion, selectivity, and yield to 1,3-PG employing Pt/WOx/Al2O3 catalysts. Reaction conditions: 220 °C, 6 MPa H2, 30 wt.% glycerol solution, 5 h. Reproduced from [104] with permission from the American Chemical Society (ACS). The results showed that even regenerating the catalyst in H 2 flow at 300 • C after each use, the yield to 1,3-PG decreased from 31 to 19% after the fourth reaction cycle. The characterization results indicated the formation of carbon deposits on the catalytic surface, which would cause the deactivation phenomenon [104].
In contrast to the previously obtained results, Yan et al. tested the Pt/Al-WO x catalysts in batch reactors at 140 • C and 3 MPa H 2 using 5 wt.% glycerol solutions and observed no apparent catalyst deactivation ( Figure 22). Reactions 2022, 3, FOR PEER REVIEW 40 The results showed that even regenerating the catalyst in H2 flow at 300 °C after each use, the yield to 1,3-PG decreased from 31 to 19% after the fourth reaction cycle. The characterization results indicated the formation of carbon deposits on the catalytic surface, which would cause the deactivation phenomenon [104].
In contrast to the previously obtained results, Yan et al. tested the Pt/Al-WOx catalysts in batch reactors at 140 °C and 3 MPa H2 using 5 wt.% glycerol solutions and observed no apparent catalyst deactivation ( Figure 22). The glycerol conversion and selectivity to 1,3-PG were maintained after four cycles of 12 h reaction each [98]. The addition of alumina as a promoter improved the catalytic properties. The characterization revealed that the high activity and selectivity of the Pt/Al-WOx catalyst was a result of the increased amount of oxygen vacancies in the in situ generation of WOx species, which enhanced the selective adsorption of glycerol and in situ Brønsted acid sites for the selective activation of secondary C-O bonds, which greatly increases the hydrogenolysis yield.
For the Pt-Au/WOx/Al2O3 catalysts, which became even more active by modification of the metal phase, Wang et al. also observed a drop in conversion from 41 to 35% after four reaction cycles of 12 h each at 180 °C and 5 MPa H2 using 10 wt.% glycerol solutions. The results showed that the introduction of a very small amount of Au (up to 0.1 wt.%) into the Pt/W/Al catalyst led to an increase in the yield to 1,3-PG ( Figure 23). The glycerol conversion and selectivity to 1,3-PG were maintained after four cycles of 12 h reaction each [98]. The addition of alumina as a promoter improved the catalytic properties. The characterization revealed that the high activity and selectivity of the Pt/Al-WO x catalyst was a result of the increased amount of oxygen vacancies in the in situ generation of WO x species, which enhanced the selective adsorption of glycerol and in situ Brønsted acid sites for the selective activation of secondary C-O bonds, which greatly increases the hydrogenolysis yield.
For the Pt-Au/WO x /Al 2 O 3 catalysts, which became even more active by modification of the metal phase, Wang et al. also observed a drop in conversion from 41 to 35% after four reaction cycles of 12 h each at 180 • C and 5 MPa H 2 using 10 wt.% glycerol solutions.
The results showed that the introduction of a very small amount of Au (up to 0.1 wt.%) into the Pt/W/Al catalyst led to an increase in the yield to 1,3-PG ( Figure 23).
properties. The characterization revealed that the high activity and selectivity of the Pt/Al-WOx catalyst was a result of the increased amount of oxygen vacancies in the in situ generation of WOx species, which enhanced the selective adsorption of glycerol and in situ Brønsted acid sites for the selective activation of secondary C-O bonds, which greatly increases the hydrogenolysis yield.
For the Pt-Au/WOx/Al2O3 catalysts, which became even more active by modification of the metal phase, Wang et al. also observed a drop in conversion from 41 to 35% after four reaction cycles of 12 h each at 180 °C and 5 MPa H2 using 10 wt.% glycerol solutions. The results showed that the introduction of a very small amount of Au (up to 0.1 wt.%) into the Pt/W/Al catalyst led to an increase in the yield to 1,3-PG ( Figure 23).  Structural characterization reveals that Au likely exists as single atoms and binds to WO x , forming Au-WO x clusters that interact with Pt at the periphery of the Pt nanoparticles. The presence of Au alleviates the Pt-WOx SMSI by electron transfer from W to Au, thus allowing a more exposed Pt surface, hence enhancing the H spillover ability, and the latter is positively related to the formation of strong Brønsted acid sites responsible for the attack of the secondary C-O bond of glycerol.
The selectivity to 1,3-PG remains practically constant after four reaction cycles (Figure 24), and the small decrease in activity can be attributed to a joint effect of Pt particle agglomeration with the formation of coke deposits [115].
Reactions 2022, 3, FOR PEER REVIEW 41 Structural characterization reveals that Au likely exists as single atoms and binds to WOx, forming Au-WOx clusters that interact with Pt at the periphery of the Pt nanoparticles. The presence of Au alleviates the Pt-WOx SMSI by electron transfer from W to Au, thus allowing a more exposed Pt surface, hence enhancing the H spillover ability, and the latter is positively related to the formation of strong Brønsted acid sites responsible for the attack of the secondary C-O bond of glycerol.
The selectivity to 1,3-PG remains practically constant after four reaction cycles (Figure 24), and the small decrease in activity can be attributed to a joint effect of Pt particle agglomeration with the formation of coke deposits [115]. Reproduced from [115] with permission from the American Chemical Society (ACS).

Conclusions, Future Work, and Perspectives
The synthesis of 1,3-PG from glycerol obtained as a by-product in biodiesel production offers an attractive alternative route to those currently implemented in the petrochemical industry.
However, the steric hindrance of the secondary -OH group in the glycerol molecule makes hydrogenolysis to 1,3-PG more difficult than the hydrogenolysis of glycerol to 1,2-PG. Of the four reaction mechanisms leading to the formation of 1,3-PG by glycerol hydrogenolysis, the dehydration-hydrogenation mechanism and the direct mechanism of hydrogenolysis have been the most widely reported. Reproduced from [115] with permission from the American Chemical Society (ACS).

Conclusions, Future Work, and Perspectives
The synthesis of 1,3-PG from glycerol obtained as a by-product in biodiesel production offers an attractive alternative route to those currently implemented in the petrochemical industry.
However, the steric hindrance of the secondary -OH group in the glycerol molecule makes hydrogenolysis to 1,3-PG more difficult than the hydrogenolysis of glycerol to 1,2-PG. Of the four reaction mechanisms leading to the formation of 1,3-PG by glycerol hydrogenolysis, the dehydration-hydrogenation mechanism and the direct mechanism of hydrogenolysis have been the most widely reported.
The glycerol hydrogenolysis reaction can be carried out in batch reactors or flow reactors operating in the liquid phase or in flow reactors operating in the vapor phase. Experience shows that the highest yields to 1,3-PG are obtained by operating in the liquid phase, with batch reactors and flow reactors being the best alternatives. The selection of the optimum operating condition in these reactors involves determining the temperature, pressure, reaction time, space velocity, H 2 flow, and concentration of glycerol, water, and eventually some solvent to maximize the yield at 1,3-PG, as appropriate.
Glycerol hydrogenolysis is a catalyst structure-sensitive reaction, requiring a metal for the activation of the H 2 molecule coupled with an oxyphilic promoter for the activation of the secondary hydroxyl group of the glycerol molecule. Pt-based catalysts supported on WO x and heteropolyacids, and other systems based on Ir, Rh modified with ReO x , in addition to some catalysts based on non-noble metals, have been reported. In the case of catalysts based on Pt and heteropolyacids, the presence of Brønsted acid sites favors the formation of the intermediate 3-HPA, which is then hydrogenated on the metal sites to form 1,3-PG, giving priority in this case to the dehydration-hydrogenation mechanism. On the other hand, in the catalytic systems based on WO x and ReO x , the generation of H + and H − species intervene in the formation of alkoxide-type intermediates that favor the formation of 1,3-PG by means of a direct mechanism of hydrogenolysis.
Of all the catalysts studied thus far, it is possible to establish the following order of activity, taking into account the yields at 1,3-PG at the optimum reaction condition: Pt/WO x /AlOOH (66%) > Pt/W-MCFs (65%) > Pt/WO x /SBA-15 (61%) > Pt/WO x /ZrO 2 (49%) > Ir-ReO x /SiO 2 (38%) The results indicate that it is possible to produce 1,3-PG with yields of 38-66% in the range of 120-180 • C and 4-8 MPa H 2 operating in batch reactors in the liquid phase condition.
The stability analysis indicates that the hydrothermal condition of the reaction affects the structural stability of the supports (Al 2 O 3 , ZrO 2 , SiO 2 ), being required for its stabilization by some thermal method or the incorporation of other elements (Li, Mg). On the other hand, the sintering by agglomeration of Pt metal particles, the loss by leaching of W species, and the formation of carbonaceous deposits have been identified as the main mechanisms of deactivation of the catalysts studied.
The yields at 1,3-PG obtained with current catalytic systems (38-66%) suggest that the formation of 1,3-PG is a disadvantage with respect to other hydrogenolysis products such as 1,2-PG. From a thermodynamic point of view, it has been reported that the formation of 3-HPA, intermediate in the formation of 1,3-PG, is thermodynamically unstable but kinetically stable, while for acetol, an intermediate in the formation of 1,2-PG, exactly the opposite occurs. This aspect is key for the design of a suitable catalyst, so it is necessary to have thermodynamic studies in the liquid phase to establish the values of conversion and selectivity to 1,3-PG to determine whether its formation is limited by equilibrium, considering, in addition, other side products in the hydrogenolysis reaction.
The design of better catalysts should be based on finishing the right set of textural, acidic, and structural properties that allow for the values reported by thermodynamics to be reached. In this sense, catalytic systems with adequate density and type of acid site, high dispersion of the metal phase, and good stability in the hydrothermal condition of the reaction are required. Due to the costs of metal phases based on noble metals, the optimization of the metal content should be a factor to be taken into account in the design of these catalysts, or their partial replacement by other less expensive metal phases.
In order to design reactors with industrial application, kinetic studies should be carried out for the formation of 1,3-PG, involving the possibility of using crude glycerol as feedstock. In addition, techno-economic and environmental impact studies would be useful for the analysis of the process in the framework of the design of glycerol biorefineries.