WO2024156344A1

WO2024156344A1 - Liquid phase process for preparing acrylic acid from glycerol

Info

Publication number: WO2024156344A1
Application number: PCT/EP2023/051748
Authority: WO
Inventors: Jakob Albert; Anna BUKOWSKI; Jan-Dominik KRUEGER
Original assignee: Universität Hamburg
Priority date: 2023-01-25
Filing date: 2023-01-25
Publication date: 2024-08-02

Abstract

The invention relates to a process for preparing acrylic acid from glycerol comprising (i) dehydration of glycerol to acrolein, and (ii) oxidation of acrolein to acrylic acid, wherein (i) and (ii) are performed in a liquid phase in the presence of a homogeneous catalyst and a solvent. The invention further relates to a composition comprising or consisting of a reaction product prepared by the process for preparing acrylic acid from glycerol.

Description

LIQUID PHASE PROCESS FOR PREPARING ACRYLIC ACID FROM GLYCEROL

DESCRIPTION

The invention is in the field of chemistry, in particular acrylic acid synthesis.

The invention relates to a process for preparing acrylic acid from glycerol comprising (i) dehydration of glycerol to acrolein, and (ii) oxidation of acrolein to acrylic acid, wherein (i) and (ii) are performed in a liquid phase in the presence of a homogeneous catalyst and a solvent.

The invention further relates to a process for preparing acrylic acid, wherein the homogeneous catalyst is a polyoxometalate.

The invention further relates to a composition comprising or consisting of a reaction product prepared by the process for preparing acrylic acid from glycerol.

BACKGROUND OF THE INVENTION

Acrylic acid (prop-2-enoic acid) is a high-demand product with various uses in industry and daily life, with a market volume of 5 billion EUR and 7.5 million tons annually in 2019 (“Global Acrylic Acid Market - Trends, COVID-19 impact and Growth Forecasts to 2029”, Chemlntel360 2022). It is used for the synthesis of various homopolymers and copolymers used in surface coatings, paint, and adhesives, such as pressure-sensitive adhesives, for the production of superabsorbers and for organic synthesis.

Acrylic acid is usually produced by steam cracking of petroleum and subsequent oxidation of propene to acrylic acid. The process is a two-stage heterogenous oxidation of propylene with bismuth-molybdate and molybdenum-vanadium catalysts in the vapor-phase. Thereby, propylene needs to be produced in an energy intensive steam-cracker process (Kampe et al., 2007). As the process is based on fossil resources it is highly energy and cost intense. In the context of the climate change there is now an increased need for sustainable chemicals produced from renewable resources minimizing their ecological footprint.

Several approaches have been made to develop sustainable processes for the production of acrylic acid from renewable resources. One approach is the NADA (Nucleophile Assisted Dehydration to Acrylates) process for lactic acid in ionic liquids e.g., [PBu4]Br (Collias et al., 2019). Another approach is the fermentation of glucose and a following dehydration of 3- hydroxypropionic acid to acrylic acid (“From the Sugar Platform to Biofuels and Biochemicals", Final Report for the European Commission, Contract No. ENER/C2/423-2012/SI2.673791 , 2015).

Another promising renewable resource available in large quantities is glycerol. Glycerol is a major by-product from the production of biodiesel, where e.g., rape seed oil is used as a renewable resource and converted to fatty acid methyl esters. As biodiesel production increased enormously within the last decades, glycerol is a highly available commodity. It is used in the pharmaceutical, cosmetics and food industries. Further glycerol can be used e.g., for the synthesis of acrolein, that can be further processed to acrylic acid (Talebian-Kiakalaieh et al., 2014). Hereby, different approaches have been developed for the conversion of glycerol to acrylic acid, mostly based on a two-step conversion comprising dehydration of glycerol to acrolein in a first step, followed by oxidation of acrolein to acrylic acid using molecular oxygen in a second step.

Approaches targeting the first step of dehydration of glycerol to acrolein include homogenous catalyzed reactions in sub- and supercritical water with zinc sulfate (Ott et al., 2006), in sulfuric acid (Groll et al., 1936), or with heterogenous catalyst as heteropoly acids, zeolites, metal oxides and phosphate catalysts (Sun et al., 2017). Such an approach for the conversion of glycerol to acrolein on a solid catalyst, i.e., in a heterogeneous reaction system, is disclosed in US 8,252,960 B2.

The oxidation step of acrolein to acrylic acid is mainly accomplished by heterogeneous vaporphase oxidation over vanadium-molybdenum, cobalt-molybdenum, vanadium-antimony catalysts (Kampe et al., 2007; Chen et al., 2013; Nojiri et al., 1995; Petzold et al., 2015) or carbon nanotubes (Frank et al., 2011 ; Zhong et al., 2014). Here, numerous promoters as tungsten, copper, manganese, iron, antimony, chrome, or strontium and supports like alumina are used. The standard conditions are temperatures between 200 to 350 °C.

Approaches in which glycerol is converted in a two-stage heterogeneous vapor-phase process combining a first dehydration reaction of glycerol to acrolein with a second oxidation reaction of acrolein to acrylic acid are disclosed in US 8,212,070 B2 and US 2011/0112330 A1 .

Another possibility is the direct oxidehydration of glycerol to acrylic acid with acrolein as intermediate in a single-step process using vapor-phase reactions over stacked catalysts, usually requiring reaction times of up to 37 h and temperatures of 250 - 350 °C (Sun et al., 2017). Such a process for manufacturing acrylic acid from glycerol in a single step-approach is disclosed in US 7,910,771 B2.

These two-step and single step approaches are usually performed under high pressure and at high temperatures and are thus highly energy- and cost-intensive. An aim to reduce the need of energy is the liquid-phase oxidation of acrolein to acrylic acid. However, such approaches, e.g., liquid phase oxidation reaction over Cu/Si02-Mn02 in acetonitrile, to date are usually not as effective as vapor-phase reactions and further often require the use of organic or corrosive solvents or harmful oxidants in stoichiometric amounts as potassium permanganate, chromium trioxide and potassium periodate in hazardous solvents (Sarkar et al., 2014; Yu et al., 2017).

Despite the recent developments in using glycerol as sustainable resource for acrylic acid production, the approaches known in the prior art are highly energy- and cost-intensive processes performed at high temperature and under high pressure.

Furthermore, the methods known in the prior art often require the use of harmful and corrosive organic solvents and oxidants.

There is thus an urgent need for environmentally friendly, cost-effective methods for producing acrylic acid from a sustainable resource available in large quantities, which do not require high amounts of energy or harmful reactants and solvents. SUMMARY OF THE INVENTION

In light of the prior art the technical problem underlying the present invention is the provision of improved or alternative means for the production of acrylic acid, that are energy-efficient, environmentally friendly and cost-effective.

A task of the invention may also be viewed as the provision of improved or alternative means for the production of acrylic acid from glycerol.

Another problem underlying the present invention was the provision of an efficient liquid phase process for the synthesis of acrylic acid from glycerol, avoiding the use of hazardous solvents and oxidants.

Another problem underlying the present invention was the provision of a catalyst that is suitable for use in a liquid phase process for producing acrylic acid from glycerol.

These problems are solved by the features of the independent claims. Preferred embodiments of the present invention are provided by the dependent claims.

The invention therefore relates to a process for preparing acrylic acid from glycerol comprising (i) dehydration of glycerol to acrolein, and (ii) oxidation of acrolein to acrylic acid, wherein (i) and (ii) are performed in a liquid phase in the presence of a homogeneous catalyst and a solvent.

Surprisingly the process of the present invention enables the production of acrylic acid from glycerol in a two-step liquid phase process. The process is performed in the presence of a highly reactive homogeneous catalyst in the liquid phase, allowing to perform the process at low temperatures and pressure. In comparison, the heterogeneous vapor-phase processes of the prior art usually require high pressures and temperatures. The liquid phase process of the present invention is thus advantageously an energy- and cost-efficient process in comparison to the prior art.

Further, the process of the present invention enables efficient production of acrylic acid from glycerol. Glycerol is a sustainable resource available in large quantities and produced from renewable resources. The process of the present invention is thus an environmentally friendly and energy-efficient approach for the production of acrylic acid.

In one embodiment the homogeneous catalyst is a polyoxometalate.

Advantageously, the polyoxometalates of the present invention can be dissolved in the liquid phase and thus enable homogeneous catalysis within the liquid phase process. Homogeneous catalysts are highly reactive and enable the performance of the liquid phase process at lower temperatures and pressures in comparison to heterogeneous reaction systems of the prior art.

In one embodiment, the dehydration of glycerol (i) is performed in the presence of a polyoxometalate having a Keggin-structure H_a[XMi2C>4o], wherein X is a hetero atom (preferably Si or P), M is one or more different metals, preferably one, two or three different metals, wherein at least one metal is a transition metal, preferably Mo, W, or V, and a is 3+n, wherein n is the charge of [XMi204o]ⁿ‘. In one embodiment the dehydration of glycerol (i) is performed in the presence of a polyoxometalate H4SiWi204o, H3PW12O40 or H3PM012O40.

In one embodiment the dehydration of glycerol (i) is performed in the presence of a glycerin ester as the solvent, preferably glycerol monostearate.

It is advantageous over the prior art that the dehydration of glycerol to acrolein can be performed within the process of the present invention in nonhazardous solvents such as glycerol monostearate commonly used in pharmaceutical and food industry.

In one embodiment the dehydration of glycerol (i) is performed at 150 °C to 300 °C. In one embodiment the dehydration of glycerol (i) is performed at 150 to 300 °C, such as 150 °C, 160 °C, 170 °C, 180 °C, 190 °C, 200 °C, 210 °C, 220 °C, 230 °C, 240 °C, 250 °C, 260 °C, 270 °C, 280 °C, 290 °C or 300 °C. In one embodiment the dehydration of glycerol (i) is performed at 170 °C to 230 °C.

In one embodiment the process comprises additionally a distillation of acrolein into a separate reaction vessel subsequent to the dehydration of glycerol (i).

In one embodiment the process comprises addition of an inhibitor of the polymerization of acrolein and/or acrylic acid subsequent to the dehydration of glycerol (i) and/or the oxidation of acrolein (ii).

In one embodiment the inhibitor of the polymerization of acrolein and/or acrylic acid is selected from phenol, thiazine, derivatives and/or mixtures thereof, preferably from hydroquinone, 4- methoxyphenol, phenothiazine and/or mixtures thereof.

A challenge in liquid phase reactions, in particular liquid phase oxidations of acrolein, is the high reactivity and polymerizing behavior of acrolein and the reaction product acrylic acid, resulting in decreased yields of acrylic acid and thus reduced efficiency of the production process. The addition of polymerization inhibitors advantageously prevents polymerization of acrolein and acrylic acid and thus results in higher yield and efficiency of the reaction.

In one embodiment the oxidation of acrolein (ii) is performed in the presence of a polyoxometalate having a Keggin-structure K_a[XMi2C>4o], a Lindqvist-structure Ka[MeOi9] or a Wells-Dawson structure Ka[X2MisO62], wherein K is H, an alkali metal or ammonium, X is a hetero atom (preferably Si or P), M is one or more different metals, preferably one, two or three different metals, wherein at least one metal is a transition metal, preferably Mo, W, V, Ni, Nb, Mn, Co or Cu, and a is 3+n, wherein n is the charge of [XMi2C>4o]ⁿ', [MeO-ig]"' or [X2Mi8O62]ⁿ‘.

In one embodiment the oxidation of acrolein (ii) is performed in the presence of a polyoxometalate H15PC03M09O40, HnPMnNiMoio04o or K10 2W17C0O62.

In one embodiment the oxidation of acrolein (ii) is performed in the presence of oxygen.

In one embodiment the oxygen is in the form of pure oxygen, air or a mixture of gases containing oxygen.

In one embodiment the oxidation of acrolein (ii) is performed at a partial oxygen pressure of 5 to 21 bar, preferably 15 bar. In one embodiment the oxidation of acrolein (ii) is performed at a partial oxygen pressure of 5 to 21 bar such as 5 bar, 6 bar, 7 bar, 8 bar, 9 bar, 10 bar, 11 bar, 12 bar, 13 bar, 14 bar, 15 bar, 16 bar, 17 bar, 18 bar, 19 bar, 20 bar or 21 bar, preferably 15 bar. In one embodiment the oxidation of acrolein (ii) is performed at a partial oxygen pressure of 5 to 21 bar and a total reaction pressure of 30 bar. In one embodiment the oxidation of acrolein (ii) is performed at a reaction pressure of 5 to 21 bar and a reaction temperature of 180 °C. In one embodiment the oxidation of acrolein (ii) is performed at a reaction pressure of 5 to 21 bar, a total reaction pressure of 30 bar and a reaction temperature of 180 °C.

Surprisingly the liquid phase process of the present invention can be performed by using oxygen as a mild oxidant, thus avoiding the use of hazardous oxidants often required in liquid phase processes of the prior art such as potassium permanganate, chromium trioxide and potassium periodate. The process of the present invention is thus a safe and environmentally friendly process for the production of acrylic acid from glycerol.

In one embodiment the oxidation of acrolein (ii) is performed at 100 °C to 200 °C and a pressure of 1 to 80 bar, preferably at 160 °C to 200 °C and a pressure of 30 to 50 bar. In one embodiment the oxidation of acrolein (ii) is performed at 100 to 200 °C, such as 100 °C, 110 °C, 120 °C, 130 °C, 140 °C, 150 °C, 160 °C, 170 °C, 180 °C, 188 °C, 190 °C or 200 °C, and a pressure of 1 to 80 bar, such as 1 bar, 5 bar, 10 bar, 15 bar, 20 bar, 25 bar, 30 bar, 35 bar, 40 bar, 45 bar, 50 bar, 55 bar, 60 bar, 65 bar, 70 bar, 75 bar or 80 bar, preferably at 160 °C to 200 °C and a pressure of 30 to 50 bar.

In one embodiment the oxidation of acrolein (ii) is performed at 160 °C to 200 °C in a reaction time of 80 to 240 min. In one embodiment the oxidation of acrolein (ii) is performed at 160 °C to 200 °C in a reaction time of 80 to 240 min, such as 80 min, 85 min, 88 min, 90 min, 95 min, 100 min, 105 min, 110 min, 115 min, 120 min, 121 min, 125 min, 130 min, 135 min, 140 min, 145 min,

150 min, 155 min, 156 min, 160 min, 165 min, 170 min, 180 min, 181 min, 182 min, 185 min, 190 min, 195 min, 200 min, 205 min, 210 min, 215 min, 220 min, 225 min, 230 min, 235 min or 240 min.

In one embodiment the oxidation of acrolein (ii) is performed at 180 °C to 200 ° in a reaction time of 30 to 240 min. In one embodiment the oxidation of acrolein (ii) is performed at 180 °C to 200 °C in a reaction time of 30 to 240 min, such as 30 min, 35 min, 40 min, 45 min, 50 min, 55 min, 60 min, 65 min, 70 min, 75 min, 80 min, 85 min, 88 min, 90 min, 95 min, 100 min, 105 min, 110 min, 115 min, 120 min, 121 min, 125 min, 130 min, 135 min, 140 min, 145 min, 150 min, 155 min,

156 min, 160 min, 165 min, 170 min, 180 min, 181 min, 182 min, 185 min, 190 min, 195 min, 200 min, 205 min, 210 min, 215 min, 220 min, 225 min, 230 min, 235 min or 240 min.

In one embodiment the oxidation of acrolein (ii) is performed in the presence of water as the solvent.

Within the process of the present invention surprisingly no organic solvents are required for the dehydration of glycerol to acrolein as wells as for the oxidation of acrolein to acrylic acid. In contrast to the prior art, the process of the present invention thus advantageously avoids the use of organic and hazardous solvents, thereby providing a safe and sustainable method for the production of acrylic acid. It was particularly surprising that the oxidation of acrolein to acrylic acid using oxygen as oxidant can be performed in a liquid phase process in water. Oxygen is known to have a low solubility in water, which can potentially result in lower efficiency of the oxidation reaction of acrolein. Advantageously under the conditions of the process of the present invention the oxidation of acrolein to acrylic acid can be efficiently performed in water using oxygen as mild antioxidant.

In one embodiment acrolein is in the form of an aqueous solution with a concentration of < 20 vol.%, preferably with a concentration of > 10 vol.%. In one embodiment acrolein is in the form of an aqueous solution with a concentration of < 20 vol.%, such as 19 vol.%, 18 vol.%, 17 vol.%, 16 vol.%, 15 vol.%, 14 vol.%, 13 vol.%, 12 vol.%, 11 vol.%, 10 vol.%, 9 vol.%, 8 vol.%, 7 vol.%, 6 vol.%, 5 vol.%, 4 vol.%, 3 vol.%, 2 vol.% or 1 vol.%, preferably with a concentration of > 10 vol.%.

In one embodiment the oxidation of acrolein (ii) is performed at a mass ratio of catalyst to acrolein of 1 :2 to 4:1 , preferably at a mass ratio 1 .5:1 . In one embodiment the oxidation of acrolein (ii) is performed at a mass ratio of catalyst to acrolein of 1 :2 to 4:1 , such as 1 :2, 1 :1 .5, 1 :1 , 1 .5:1 , 2:1 , 2.5:1 , 3:1 , 3.5:1 or 4:1 , preferably at a mass ratio of 1.5:1 .

In one embodiment the oxidation of acrolein (ii) is performed in a stirred tank reactor or a liquid phase Berty-reactor. In one embodiment the oxidation of acrolein (ii) is performed in a stirred tank reactor in batch modus, semi-batch modus or continuous modus.

In one embodiment the yield of acrylic acid from the oxidation of acrolein to acrylic acid (ii) is > 7 mol%, preferably > 15 mol%, more preferably > 20 mol% and even more preferably > 25 mol%. In one embodiment the yield of formic acid from the oxidation of acrolein to acrylic acid (ii) is < 35 mol%, preferably < 30 mol%, more preferably < 25 mol%. In one embodiment the yield of acrylic acid from the oxidation of acrolein to acrylic acid (ii) is 7 mol%, 10 mol%, 15 mol%, 20 mol%, 25 mol%, 30 mol%, 35 mol%, 40 mol%, 45 mol%, 50 mol%, 55 mol%, 60 mol%, 65 mol%, 70 mol%, 75 mol%, 80 mol%, 85 mol%, 90 mol%, 95 mol% or 100 mol%. In one embodiment the yield of formic acid from the oxidation of acrolein to acrylic acid (ii) is 35 mol%, 30 mol%, 25 mol%, 20 mol%, 15 mol%, 10 mol%, 5 mol% or O mol%.

In one embodiment the conversion of acrolein in the oxidation of acrolein to acrylic acid (ii) is > 60%, preferably > 70%, more preferably > 80%, such as 60%, 65%, 70%, 75%, 80%, 85%, 90%, 95% or 100%.

In one embodiment the liquid phase selectivity of acrylic acid in the oxidation of acrolein to acrylic acid (ii) is > 15%, preferably > 20%, more preferably > 30%, such as 15%, 20%, 25%, 30%, 35%, 40%, 45%, 50%, 55%, 60%, 65%, 70%, 75%, 80%, 85%, 90%, 95% or 100%.

In one embodiment the yield of acrylic acid from the oxidation of acrolein to acrylic acid (ii) is > 7 mol%, preferably > 15 mol%, more preferably > 20 mol% and even more preferably > 25 mol% and the liquid phase selectivity of acrylic acid in the oxidation of acrolein to acrylic acid (ii) is > 15%, preferably > 20%, more preferably > 30%.

In one aspect the invention relates to a reaction product prepared by the process for preparing glycerol.

In one embodiment the reaction product comprises acrylic acid and one or more of formic acid, acetic acid, glycolic acid and/or glycerol. In one embodiment the reaction product comprises acrylic acid in an amount of > 7 mol%, such as 7 mol%, 10 mol%, 15 mol%, 20 mol%, 25 mol%, 30 mol%, 40 mol%, 45 mol%, 50 mol%, 55 mol%, 60 mol%, 65 mol%, 70 mol%, 75 mol%, 80 mol%, 85 mol%, 90 mol%, 95 mol% or 100 mol%, and formic acid in an amount of < 50 mol%, such as 50 mol%, 45 mol%, 40 mol%, 35 mol%, 30 mol%, 25 mol%, 20 mol%, 15 mol%, 10 mol%, 5 mol% or O mol%, preferably < 40 mol%. In a further embodiment the reaction product comprises acrylic acid in an amount of > 15 mol%, more preferably > 20 mol%.

The features above with respect to the process for preparing acrylic acid from glycerol have a structural and functional outcome on the reaction product of the present invention, such that the reaction product can in some embodiments be described by features of or derived from the process of preparing, and vice versa.

DETAILED DESCRIPTION OF THE INVENTION

Acrylic acid (C3H4O2, CAS 79-10-7) is an unsaturated carboxylic acid. It is a reactive liquid with a melting point of 13.5 °C and a boiling point of 139 °C, which polymerizes spontaneously and is miscible with water, alcohols, ethers and chloroform. Due to its high reactivity, acrylic acid is mostly stabilized using polymerization inhibitors such as methyl hydroquinone (MEHQ). Acrylic acid is used for the synthesis of various homopolymers and copolymers used in surface coatings, paint, adhesives such as pressure-sensitive adhesives, for the production of superabsorbers and for organic synthesis. Acrylic acid is commonly produced by oxidation of propylene, which is a byproduct of the production of ethylene and gasoline. The term “acrylic acid” includes the acrylate ion and salt forms of acrylic acid.

The chemical structure of acrylic acid is:

Glycerol, also termed “glycerine” or “trihydroxypropane” (C3H8O3, CAS 56-81-5), is a colorless, odorless, viscous, non-toxic liquid. Glycerol can be obtained from plant and animal sources, where it occurs mainly in triglycerides. The hydrolysis, transesterification or saponification of these triglycerides produces glycerol. Sources for the production of glycerol include without limitation soybeans, palm, vegetable oils, rapeseed, jatropha, mahua, mustard, flax, sunflower, hemp, field pennycress, Pongamia pinnata and algae. Glycerol is further a byproduct of biofuel production such as of the production of biodiesel, usually obtained by transesterification of vegetable oil. According to the present invention the purity of glycerol is preferably 80 wt% or more, preferably 90 wt% or more, and more preferably 95 wt% or more, in order to reduce the production of reaction byproducts.

The chemical structure of glycerol is:

Acrolein, also termed “2-propenal” (C3H4O, CAS 107-02-8), is an unsaturated aldehyde with an oil-like consistency. It is present in cooked foods and in the environment due to combustion of fuels or smoking of tobacco products (Stevens et al., 2008). Acrolein is a colorless liquid with a low boiling point of 52 °C and a very high vapor pressure and is derived by decomposition of glycerol. In the industry acrolein is used as intermediate in the synthesis of DL-Methionine, an important compound feed ingredient, as biocide and for acrylic acid. As acrolein is a a, p- unsaturated aldehyde it possesses a conjugated ir-electron system with overlapping IT orbitals allowing the electrons to move freely within the orbitals. The a, p-unsaturated system is also called MICHAEL-system. Due to the C=C- and the C=0-double bond acrolein is a very reactive compound. The reactivity of acrolein is further increased by the electron withdrawing effect of the carbonyl group, resulting in a preference for an electrophilic rather than a nucleophilic attack. Acrolein can spontaneously polymerize or react with itself in a DIELS-ALDER-reaction. Further, acrolein can react in 1 ,2- and 1 ,4-additions and with CH-acidic chemicals in a MICHAEL-addition. Due to this reactivity for storage of pure acrolein usually polymerization inhibitors are added and the mixture is stored under nitrogen.

The chemical structure of acrolein is:

For the conversion of glycerol (1) to acrylic acid (3) according to the present invention, glycerol (1) is dehydrated to acrolein (2) followed by selective oxidation of the carbonyl group of acrolein to acrylic acid (3) according to the following reaction scheme:

Possible side reactions of the oxidation of acrolein to acrylic acid include C-C-cleavage reactions of the C=C double bond. These side reactions include double bond oxidation, oxidative C-C cleavage and the thermal degradation resulting in acids such as formic acid, acetic acid and glycolic acid, aldehydes, water and carbon oxides such as carbon monoxide and carbon dioxide as reaction products (Thanasilp et al., 2013; Miller et al., 2018). C1-C2 side products (CO2, CO, acetylene, acetic acid, ethylene, formic acid) are formed via C-C bond scission of a single or a double C-C bond of acrolein and/or acrylic acid. The C4-C7 products are formed by addition of one or two acrolein and/or acrylic acid molecules to a C2 surface species coupled with decarbonylation or decarboxylation reactions (furan, furfuraldehyde, butadiene, benzaldehyde benzene, butenone) (Miller et al. 2018).

The term “dehydration” or “dehydration reaction” refers to a chemical reaction that involves the loss of water from a reacting molecule such as glycerol. According to the present invention the dehydration of glycerol is performed in the presence of an acid also termed “acidic compound”, preferably a Bransted-acidic compound. According to the present invention a Bransted acid or a Bransted-acidic compound also termed “proton donator” is a compound that is capable of transferring a proton (H⁺) to a Bransted base also termed “proton acceptor”, which is a compound capable of accepting a proton.

In embodiments the dehydration of glycerol is performed in the presence of a polyoxometalate which is Bransted-acidic. Bransted-acidic polyoxometalates are according to the present invention polyoxometalates having a Keggin-structure H_a[XMi2C>4o], wherein X is a hetero atom (preferably Si or P), M is one or more different metals, preferably one, two or three different metals, wherein at least one metal is a transition metal, preferably Mo, W, or V, and a is 3+n, wherein n is the charge of [XMi204o]ⁿ‘. In embodiments Bransted-acidic polyoxometalates include without limitation H4SiWi204o, H3 W12O40 and H3 M012O40.

The term “oxidation” or “oxidation reaction” refers to a chemical reaction in which one molecule termed “electron donor” donates electrons to another molecule termed “electron acceptor” or “oxidant”. The oxidant thereby is reduced by the acceptance of the electron. An oxidation reaction is therefore always associated with a reduction reaction. Oxidation and reduction are considered as partial reactions of a redox reaction. According to the present invention acrolein as electron donator is oxidized by oxygen as electron acceptor. In embodiments the reaction is performed using pure oxygen, air or a mixture of gases containing oxygen. In embodiments the mixture of gases may comprise besides oxygen other gases such as without limitation nitrogen, argon, carbon dioxide and/or sulfur dioxide.

The term "catalyst" refers to a substance that increases the reaction rate of a chemical reaction by lowering the activation energy of the reaction, typically without being consumed by the chemical reaction itself. When used in reference to a dehydration reaction such as the dehydration of glycerol to acrolein, the term refers to a substance that increases the reaction rate of the dehydration reaction without being consumed itself. When used in reference to an oxidation reaction such as the oxidation of acrolein to acrylic acid, the term refers to a substance that increases the reaction rate of the oxidation reaction without being consumed itself.

Depending on the phases in which catalyst and reactants are present, a catalyst is referred to as homogeneous or heterogeneous catalysts. A homogeneous catalyst is a catalyst which is present in the same phase as the reactants and the reaction takes place in. Homogeneous catalysts are often liquid and catalyze reactions in the liquid phase. A heterogeneous catalyst is a catalyst which is present in another phase as the reactants and the reaction takes place in.

Heterogeneous catalysts are usually present in the solid phase while catalyzing reactions in the liquid or gas phase. According to the present invention the catalyst is a homogeneous catalyst. Preferably the reactants and the homogeneous catalyst are present in a liquid phase.

A “polyoxometalate” is a polyatomic ion, usually an anion, comprising one or more transition metals, oxygen and optionally one or more heteroatoms. Polyoxometalates are usually formed by three or more transition metal oxyanions and bridged via oxygen atoms, thereby forming a three- dimensional network. Polyoxometalates can be classified into heteropolyanions and isopolyanions. Heteropolyanions comprise one or more heteroanions, such as a phosphate ion or a silicate ion. Isopolyanions do not comprise a heteroatom. Polyoxometalates can form Bransted- acidic heteropolyacids with hydrogen ions (protons). A variety of structures is known for polyoxometalates including without limitation the Keggin-structure (K_a[XMi2C>4o] and the corresponding anion [XMi204o]ⁿ'). the Lindqvist-structure (Ka[MeOi9] and the corresponding anion [MeO-ig]"'), the Wells-Dawson-structure (Ka[X2MisO62] and the corresponding anion [X2Mi8O62]ⁿ'). the Anderson-structure (K_a[XMeO24] and the corresponding anion [XMeC^f , the Allman-Waugh- structure (K_a[XMgO32] and the corresponding anion [XMgO32]ⁿ‘). the Weakley-Yamase-structure (K_a[XMioC>36] and the corresponding anion [XM-ioOse]"') and the Dexter-Silverton-structure (K_a[XMi₂O₄2] and the corresponding anion [XMi2O42]ⁿ')- Within these structures K ean be an alkali metal such as K, Na and Li, H or ammonium, X is a heteroatom such as a phosphate ion, silicate ion or sulfate ion, M is one or more different metals, usually a transition metal of groups 5 and 6 of the periodic system such as Mo, W, V, Ni, Nb, Mn, Co or Cu. Polyoxometalates are commonly used as catalysts for chemical reactions such as oxidation reactions, in molecular electronics such as in non-volatile storage devices also termed “flash memory devices” and in pharmaceutical industry as active ingredient in antiviral and antitumor medication.

A distillation is a thermal separation process for separation of components or substances from a liquid mixture using the boiling point difference of the mixture such as the difference in boiling point between glycerin (290 °C) and acrolein (53 °C) and may be carried out below the atmospheric pressure, atmospheric pressure or above the atmospheric pressure. Thereby, the component with the lower boiling point is separated by evaporation and subsequent condensation in a separate vessel. Distillation can be performed in batch or continuous mode. Suitable equipment and an apparatus for performing a distillation is known to a person skilled in the art and comprises at least a vessel wherein the liquid mixture is heated, a heater such as a heating plate or an oil bath or water bath, a condenser in which the heated vapor is condensed such as a Liebig condenser and a receiving vessel in which the condensed liquid is collected. Such an apparatus is exemplarily shown in figures 1 and 2.

A polymerization inhibitor also termed “inhibitor” or “stabilizer” is a substance that delays or prevents the polymerization of monomers. Polymerization inhibitors are usually substances that readily form mesomeric stabilized radicals through a transfer reaction. The resulting radicals are inert and do not react with the monomers to any appreciable extent. Commonly used polymerization inhibitors are without limitation phenols such as phenol, hydroquinone, 4- methoxyphenol, 4-nitrophenol and butyl hydroxytoluol, thiazines such as phenothiazine, p- phenylenediamine and hydroxylamines such as diethylhydroxylamine (DEHA).

REFERENCES

A. Talebian-Kiakalaieh, N. A. S. Amin, H. Hezaveh, Renew. Sustain. Energy Rev. 2014, 40, 28- 59.

P. Kampe, L. Giebeler, D. Samuelis, J. Kunert, A. Drochner, F. HaaB, A. H. Adams, J. Ott, S. Endres, G. Schimanke, T. Buhrmester, M. Martin, H. Fuess, H. Vogel, Phys. Chem. Chem. Phys. 2007, 9, 3577-3589.

D. Collians, J. Godlewski, J. Velasquez, P. Dziezok, M. Kehrer, J. Nagengast, J. Kadar, N.

Taccardi, J. Albert, P. Wasserscheid, in Am. Chem. Soc., Washington, 2019. L. Ott, M. Bicker, H. Vogel, Green Chem. 2006, 8, 214-220.

H. P. A. Groll, G. Hearne, Process of Converting a Polyhydric Alcohol to a Carbonyl Compound, 1936, 2042224.

D. Sun, Y. Yamada, S. Sato, W. Ueda, Green Chem. 2017, 19, 3186-3213.

B. Sarkar, C. Pendem, L. N. Sivakumar Konathala, R. Tiwari, T. Sasaki, R. Bal, Chem. Commun. 2014, 50, 9707-9710.

B. Frank, R. Blume, A. Rinaldi, A. Trunschke, R. Schlogl, DGMK Int. Conf. Catal. - Innov. Appl. Petrochemistry Refin. 2011 , 211-216.

B. Zhong, H. Liu, X. Gu, D. S. Su, ChemCatChem 2014, 6, 1553-1557.

C. Chen, N. Kosuke, T. Murayama, W. Ueda, ChemCatChem 2013, 5, 2869-2873.

N. Nojiri, Y. Sakai, Y. Watanabe, Catal. Rev. 1995, 37, 145-178.

T. Petzold, Funktion Und Bedeutung von Hydroxylgruppen Bei Der Mischoxidkatalysierten Selektivoxidation von Acrolein, Technische Universitat Darmstadt, 2015.

H. Yu, S. Ru, G. Dai, Y. Zhai, H. Lin, S. Han, Y. Wei, Angew. Chemie - Int. Ed. 2017, 56, 3867- 3871.

E. de las Heras, A. Zuriarrain-Ocio, J. Zuriarrain, A. Bordagaray, M. T. Duenas, I. Berregi, Foods 2020, 9, 1-10.

J. F. Stevens, C. S. Maier, Mol. Nutr. Food Res. 2008, 52, 7-25.

S. Thanasilp, J. W. Schwank, V. Meeyoo, S. Pengpanich, M. Hunsom, J. Mol. Catal. A Chem. 2013, 380, 49-56.

J. H. Miller, A. Bhan, ChemCatChem 2018, 10, 5242-5255.

FIGURES

The invention is demonstrated by way of the figures disclosed herein. The figures provide support for a detailed description of potentially preferred, non-limiting embodiments of the invention.

Fig.1 : Schematic representation of an experimental setup of a reactor for the dehydration reaction of glycerol to acrolein and subsequent distillation. The apparatus comprises a vessel 1 wherein the liquid mixture is heated, a heater with an oil bath or water bath 2, a condenser 3 in which the heated vapor is condensed, a receiving vessel 4 in which the condensed liquid comprising acrolein is collected and a temperature control unit 5.

Fig.2 : Schematic representation of a further experimental setup of a reactor for the dehydration reaction of glycerol to acrolein and subsequent distillation.

Fig.3: Piping and instrumentation flow scheme of a 10-fold oxidation plant for oxidation of acrolein to acrylic acid. The 10-fold oxidation plant comprises 10 Hastelloy C276 reactors (R1- R10), each equipped with analogue manometers (PI 1 -PI 10) and digital pressure transducers (PIR 1-PIR 10). Furthermore, each reactor is equipped with a rupture disc. The reactor head is linked with steel flex lines to the gas supply and is shut by ball valves individually. The 10-fold oxidation plant is connected to the gas supply by pressure reducers (V7, V9 and V10) and also equipped with analogue manometers (P11 , P13 and P15). The valves V1-V5 and V8 allow to set the gas phase individually (O2, He and N2) and to release reaction gases into the exhaust. The temperature is set, controlled and recorded by the HORST GmbH controller (TIC2 and TIR1). The stirring is controlled by SC1 . The external thermocouple TC1 allows to measure temperature individually.

Fig.4: Average yield of acrolein against the used catalyst wt.% referring to glycerol. Reaction parameters: 6.8 g of x wt.% polyoxometalate in glycerol solution, 68 g glyceride, 60 g absorbent (water & 2500 ppm hydroquinone), 230 °C, 1 atm, 3 h, batch mode.

Fig.5 : Results of variation of total reaction pressure. Reaction parameters: 5 mL of 0.14 mol/L acrolein (aq.), 10 mg catalyst, 15 bar C>2, fill up with N2, 180 °C, batch mode. Ni-Mn catalyst: HnPMnNiMoio04o; Co-catalyst: H15PC03M09O40. Y-achsis: yield in mol%.

Fig.6 : Results of variation of partial oxygen pressure. Reaction parameters: 5 mL of 0.14 mol/L acrolein (aq.), 10 mg catalyst, x bar C>2, fill up with N2, 180 °C, batch mode. Ni-Mn catalyst: HnPMnNiMoioC>4o; Co-catalyst: H15PC03M09O40. Y-achsis: yield in mol%.

Fig.7 : Results of variation of the substrate:catalyst ratio. Reaction parameters: 5 mL of 0.14 mol/L acrolein (aq.), variable catalyst mass, 15 bar C>2, 7 bar N2, 180 °C, batch mode. The Ni-Mn catalyst HnPMnNiMoioC>4o was used for the experiments.

Fig.8 : Results of variation of the substrate concentration. Reaction parameters: 5 mL of x vol.% acrolein (aq.), 1.5:1 substrate: catalyst ratio, 15 bar C>2, 7 bar N2, 180 °C, batch mode. The Ni-Mn catalyst HnPMnNiMoioC>4o was used for the experiments.

Fig.9: 3 -D contour plot for conversion. Red dots show response above predicted value, pink dots below predicted value. Plot received by Design Expert 11 .

Fig.10: 3-D contour plot for the acrylic acid yield. Red dots show response above predicted value, pink dots below predicted value. Plot received by Design Expert 11 .

Fig.11 : 3-D contour plot for the acrylic acid liquid phase selectivity. Black dots show response above predicted value, gray dots below predicted value. Plot received by Design Expert 11.

Fig.12: Numerical optimization graphs with the prediction for the best result for yield and selectivity flagged in each contour plot.

Fig.13: Results of variation of reaction time in the range of 10 - 121 min. Reaction parameters: 5 mL of 1 vol.% acrolein (aq.), 1.5:1 substrate:catalyst ratio, 15 bar C>2, 7 bar N2, 188 °C, batch mode.

Fig.14: Exemplarily shown ¹H-NMR spectrum for the dehydration reaction of glycerol to acrolein.

Fig.15: Exemplarily shown GC-MS chromatogram for the liquid phase analysis of the oxidation reaction of acrolein to acrylic acid. Fig.16: Exemplarily shown MS spectrum of acrylic acid received in liquid phase analysis using the GC-MS method.

Fig.17: Exemplarily shown HPLC chromatogram for the liquid phase analysis of the oxidation reaction of acrolein to acrylic acid. Reaction parameters: 5 mL of 1 vol.% acrolein (aq.), 1 :1 substrate:catalyst ratio, 17 bar C>2, 8 bar N2, 180 °C, 2h, batch mode.

EXAMPLES

The invention is demonstrated by way of the examples disclosed herein. The examples provide technical support for a detailed description of potentially preferred, non-limiting embodiments of the invention.

The dehydration experiments were executed in 100 mL three-necked glass flask connected to a distillation bridge with an absorbent flask. The reaction solution was stirred with a KPG-stirrer and the heating was provided by an external heater. The reaction temperature was measured and regulated by an inside sensor enveloped in a polyethylene case. The cooling for the distillation was provided by a Huber cryostat. The whole apparatus was encased in tin foil inhibiting the light induced polymerization of produced acrolein. The experimental setup is shown in figures 1 and 2.

Glycerol monostearate (68 g, 0.19 mol) used as solvent was filled into the reactor and melted at 80 °C while stirring with 15 rpm. The ice cooled aqueous absorbent (60 mL) with hydroquinone (2500 ppm, 0.15 g, 0.136 mmol) as stabilizer was stirred permanently during the reaction.

Parallel to the melting of the solvent a substrate-catalyst solution was prepared. For this glycerol (5.2 mL, 6.8 g, 72 mmol) was warmed up to 50 °C and the polyoxometalate silicotungstic acid (H4SiWi204o) was added as a homogeneous catalyst and dissolved under stirring. The reaction vessel was then heated up to 230 °C reaction temperature. When the reaction solution reached 170 °C the warmed-up glycerol-polyoxometalate solution was added. After 3 h reaction time the reaction vessel was cooled down and liquid samples were taken and analyzed with ¹H-NMR spectroscopy using n-propanol as internal standard.

Oxidation of acrolein to acrylic acid

The usage of high-pressure molecular oxygen as oxidant and the high vapor pressure of acrolein at elevated temperature require the use of a high-pressure screening plant. For this the 10-fold screening apparatus was used. Each experiment was conducted in a reactor made of Hastelloy C276 (reactor head and reactor) with an approximately total volume of 32 mL. The pipes, valves and fittings were produced out of stainless steel 1 .4761 and the Gaskets of PTFE. Each reactor is equipped with an analogue manometer and a digital pressure transducer. For safety reasons each reactor is equipped with a rupture disc (p_max = 90 bar ± 10 %) connected to the laboratory exhaust. Using steel flex lines linking either nitrogen or oxygen can be filled in the reactors. Therefore, each reactor head is equipped with ball valves by quick-release couplings. The piping and instrumentation flow scheme is shown in figure 3. Therefore, it is possible to set each reactors gas phase individually with different pressures of the oxidant molecular oxygen or the inert gas nitrogen. The reaction temperature was adjusted by a heating block and controlled by a HORST GmbH controller. For the oxidation reactions the catalysts were loaded into the reactors. Then an aqueous stock solution of acrolein was prepared and stirred permanently. Different concentrations were prepared by using volumetric percentages. For the mostly used 1 vol.% acrolein solution, 90 wt.% stabilized acrolein (1.0 mL, 0.84 g, 13.49 mmol, 0.14 mol/L) was dissolved in permanently stirred water (99.00 mL). From this solution 5.00 mL was filled into each reactor and each was closed with their corresponding reactor heads. The ball valve and the bursting disc were then coupled with the exhaust and the gas supply. Additionally, the digital pressure recording adapters were attached to each sensor on the reactor head. Before any experiment was conducted, first the leak tightness was proven at pressures at least 10 bar over reaction pressure. Afterwards the reactor was purged three times with pure oxygen and the desired partial oxygen pressure was adjusted using a needle valve. Finally, the reaction temperature was set. After the heating plate reached the set temperature the reaction start was rated 15 min later, to ensure complete heating of the reaction solution in the reactors. When the reaction was finished after the desired reaction time, the reactor heads were decoupled from the adapters and cooled down to room temperature with pressurized air under the fume hood. Gaseous samples were taken into evacuated gas bags and analyzed afterwards. The reactors were depressurized and opened to take liquid samples for qualification and quantification.

For the kinetic experiments the method was modified slightly. The reaction solution was prepared equally. But the heating plate was heated up to reaction temperature and in parallel the reactors were purged and pressurized outside the heating plate. After the temperature and the pressure was set, the reactors were placed into the heating plate. After 15 min the reaction start was defined. The cooling and analysis were executed as described before.

Analytical methods and calculations

Nuclear magnetic resonance spectroscopy (NMR) and quantification

Nuclear magnetic resonance spectroscopy (NMR) was used for liquid phase analysis. Hereby, ¹H-NMR spectra were used for the qualitative analysis and especially quantification for the dehydration reaction. The ¹H-NMR spectra were measured with a spectral width of 15.6 ppm with 128 scans at 400 MHz (Bruker).

D2O was used as a lock solvent and n-propanol was used as an internal standard for the acrolein concentration determination. An exemplarily ¹H-NMR spectrum is shown in figure 14. For acrolein quantification a NMR-method developed by Heras et al., (2020) was executed. A pulse sequence zgesgp was chosen for water suppression, with 4 dummy scans, 5.0 s acquisition time, 5.0 s recycle delay and a flip angle of 90 °. The acrolein quantification for the dehydration reaction was executed as follows:

The acrolein yield Y_p was calculated using the following equation:

Y = ^{Im ms mt} - 100 P I_e M_s m n_gly v_p The measured integral for a chosen peak l_m is multiplicated with the mass of standard added to sample m_s and the total mass of product solution mt and divided by the expected integral for the same peak l_e, the molar mass of latter M_s, the amount of sample added m together with the molar mass of used glycerol n_g/_yand the stochiometric factor v_p for product referred.

High-performance liquid chromato raphy (HPLC)

High-performance liquid chromatography was used for liquid phase product analysis. For this, a Shimadzu HPLC instrument equipped with an HPX-87H separation column from Biorad was used. The eluent was an aqueous sulfuric acid solution (5 mmol/L) with a flow rate of 0.5 mL/min at 40 °C and a column pressure of 69 bar. The oven (CTO-40S) and Rl-detector (RID-20A) were also from Shimadzu. The calibration of products was carried out corresponding to the concentration ranges used in the performed reactions. An exemplary HPLC-chromatogram is shown in figure 17.

Gas-chromatography coupled with Mass Spectrometry (GC-MS)

The acrolein conversion was measured using a gas-chromatograph coupled to a mass spectrometer. Here, an Agilent 5977B GC/MSD instrument equipped with an Agilent HP-5ms Ultra Inert column (30 m x 250 pm x 0.25 pm) with a flow rate of 1 .2 mL/min of Helium and a temperature range from 40 °C - 240 °C was used. The MS detector was closed from 3.60 - 4.10 min (isopropanol) and from 5.60 - 6.20 min (water) to impede a solvent overload and a rapid degradation of the detector with a total measurement time of 17.5 min. The samples were prepared using a 1 :1 solvation of the liquid phase products in a 1 :1 watenisopropanol mixture and an addition of 5.00 pL tetra hydrofuran (pure) as internal standard for quantification. The calibration of acrolein was executed corresponding to the concentration ranges used in the performed reactions. The GC-MS was also used for qualification of by-products in the liquid- and vapor-phase. Gaseous samples were injected manually directly into the instrument-inlet. The data was analyzed using the software Agilent Mass Hunter quantitative and qualitative analysis. Literature comparison was carried out using the NIST 11 database. An exemplary GC-MS- chromatogram and MS-spectrum are shown in figure 15 and figure 16.

Gas-chromatography (GC)

The vapor-phase samples produced by the 10-fold oxidation reactions were analyzed with a Varian 450-GC gas chromatograph (GC) equipped with a Shin-Carbon-ST-Column (2 m x 0.75 mm). The samples were injected trough a 250 pL sample-tube and led trough the stationary column phase with an argon gas-flow at 4.8 bar. The samples were examined with a thermal conductivity detector for carbon monoxide, carbon dioxide and oxygen content. The quantification is based on calibration data of these gaseous products and executed with the software Galaxy chromatography Data systems.

Calculations

The calculation of reaction indicator values was executed using the following equations. The conversion X of acrolein was calculated from data of the GC-MS analysis with the following equation:

X[%] = n ■ 100 e Here, n_e is the initial and n_p is the final amount of substrate.

The yields of isolated products were calculated with HPLC data by the following equation:

Y[mol%] = ■ 100 n_{p p}

Here, n_e is the initial amount of substrate and n_p is the product amount. The corresponding stochiometric factors are v_e and v_p.

Due to the fast polymerization of acrolein during the reaction and substrate preparation the selectivity of the liquid phase was determined by the following equation: s[%] = ^- - 100

2 ⁿx

Here, n_p is the amount of product and the sum of n_x is formed from all visible products.

For kinetic measurements the apparent activation energy E_a is calculated with the following equation:

E_A = m ■ (— R)

Here, m is the slope of a linear fit of data within the so-called ARRHENIUS plot and R is the gas constant.

Results

The dehydration of glycerol to acrolein was performed in glycerol monostearate as solvent with the polyoxometalate silicotungstic acid (H4SiWi204o) as a BR0NSTED acidic catalyst. Thereby a glycerol-polyoxometalate solution with different catalyst concentrations referring to glycerol was used. The highest yield of acrolein was achieved with 47 % at 3.0 wt.% and the lowest with 18 % at 2.0 wt.%. Another observation was the increasing purity of acrolein with higher amounts of catalyst. Therefore, in the following the catalyst concentration was increased further up to 5.0 wt.%, where the reaction is supposed to be heterogeneously catalyzed. The yield of acrolein was calculated as described above using ¹H-NMR spectroscopy and a known amount of the internal standard n-propanol.

The average yield of acrolein can be seen in figure 4, where it is plotted against the catalyst weight percentage referring to glycerol. The reaction yield is reaching a plateau at approximately 50 %. This can be attributed to reaching the meta stable regions in a heterogenous phase with a catalyst wt.% of 3.0 resulting in an increasing segregation. This leads to the assumption, that the segregation and thus heterogenic catalysis limits the mass transport to the catalyst and blocks its reaction with the substrate as a main factor, whereas the reaction can be sufficiently catalyzed under homogeneous conditions.

Overall, the highest yield with the best ratio of product-to-byproducts of acrolein with 47 % was achieved using silicotungstic acid as catalyst in glycerol monostearate as solvent at 230 °C with a reaction time of 3 h. The product acrolein and the side product water were separated in situ from the reaction solution using a distillation bridge. Here, an aqueous absorbent with dissolved hydroquinone as polymerization inhibitor was stirred and cooled permanently (< 5°C) to quench possible reactions.

Oxidation of acrolein to acrylic acid

Initial catalyst screening

To get an insight on the activity of different polyoxometalates used as catalyst and their suitability for selective acrolein oxidation the redox-active molar mass was calculated and normalized for each catalyst (here 0.79 pmol). The resulting yields of the main reaction products in the liquidphase namely acrylic acid, formic acid, acetic acid, glycolic acid and glycerol were analyzed by HPLC. The results are shown in table 1 .

Table 1 : Results for acrolein oxidation with various transition substituted polyoxometalates as homogeneous catalysts (yield Yxx of a reaction product). Reaction conditions: 5 mL of 0.14 mol/L acrolein (aq.), 0.44 pmol active metal, 15 bar C>2, 7 bar N2, 180 °C, batch mode.

The yield of acrylic acid is clearly improved in the presence of a polyoxometalate as catalyst. In the blank experiment a yield of 7 % acrylic acid and almost 40% formic acid yield were observed. This high yield shows that a C-C cleavage side reaction of acrolein to formic acid is the main pathway for the uncatalyzed reaction. Using catalyst improves the acrylic acid yield significantly lowering the effect of the C-C cleavage reaction. The best catalyst regarding the yield of acrylic and production of unwanted by-products is the triple substituted cobalt polyoxometalate (H7PC0M011O40, H11PC02M010O40 and H15PC03M09O40) with a total yield of 15 % of acrylic acid, followed by the nickel-manganese substituted polyoxometalate (HnPMnNiMoio04o) with a yield of 14 % acrylic acid.

For further investigations the acrolein conversion as well as the selectivity towards acrylic acid was further investigated using GC-MS, which allows acrolein quantification with an internal standard.

The catalysts H15PC03M09O40, HnPMnNiMoio04o, Ky[P2Wi7Nb2O62] and K [P2Wi7Co062] were investigated by GC-MS. All experiments were at least conducted three times and the results are shown in table 2. Table 2: Results for acrolein oxidation with various transition substituted polyoxometalates as homogeneous catalysts. Averaged results of acrolein conversion (X), acrylic acid selectivity (S) and product yields Y. Reaction conditions: 5 mL of 0.14 mol/L acrolein (aq.), 0.44 pmol active metal, 15 bar C>2, 7 bar N2, 180 °C, batch mode. Conversion received by GC-MS, yield and selectivity received by HPLC.

In all performed experiments acrolein was converted from 86 - 95 % showing a high reactivity of the chosen system. The yields for acetic acid, glycolic acid and glycerol are combined about 10 % of the overall yield for all reactions, making them minor products. The highest overall yield is received for the formic acid ranging from 17 - 39 %. This can be led back to major thermal C-C cleavage reactions. The best result for acrylic acid yield was achieved with 17 % acrylic acid yield and a selectivity of 38 % using the triple substituted cobalt polyoxometalate. Following is the nickel-manganese substituted polyoxometalate (H15PC03M09O40) with 16 % acrylic acid yield and 31 % selectivity. Overall, all used catalyst showed substantially higher acrylic acid yields and selectivity compared to the blank experiment without catalyst.

Optimization of reaction conditions

The parameters of interest are partial oxygen pressure, total reaction pressure, substrate-to- catalyst ratio, substrate concentration and the establishment of the substitution of the polyoxometalate (nickel-manganese or cobalt). For these investigations H15PC03M09O40 (cobalt; Co) and HnPMnNiMoio04o (nickel-manganese; Ni-Mn) are used as catalyst.

Reaction pressure:

The reaction pressure at elevated temperatures needs to be higher than the vapor pressure of acrolein to perform a liquid phase acrylic acid synthesis. Firstly, the influence of the total reaction pressure was investigated at 30 to 50 bar for both catalysts and blank at 180 °C reaction temperature (Figure 5). The yield of acrylic acid is in the range of 19 - 22 % for the nickel- manganese substituted catalyst (HnPMnNiMoio04o) and in the range from 17 - 19 % for the triple cobalt substituted catalyst (H15PC03M09O40) for all tested total pressures. The results show a slightly decreasing overall yield with increasing pressure. Taking further economic and environmental factors into account lower pressures are preferred and for further experiments the total reaction pressures was fixed at 30 bar.

Partial oxygen pressure: In the following the partial oxygen pressure is varied from 5 bar to 21 bar at 180 °C reaction temperature. The oxygen pressure of 21 bar means there is no additional nitrogen filled into the reactor. After the reaction temperature of 180 °C the total pressure in the reactors was 30 bar (Figure 6).

The highest acrylic acid yield is 22 %, achieved by using the nickel-manganese substituted catalyst (HnPMnNiMoio04o) at 15 bar partial pressure of oxygen. The remaining partial pressures show slightly lower yields with 19 - 20 % of acrylic acid. The acrylic acid yields using the cobalt substituted catalyst (H15PC03M09O40) are ranging from 14 - 18 %. The blank reaction shows the lowest acrylic acid yield ranging from 10 - 12 % with significantly higher yields of formic acid. These results show that the partial oxygen pressure is not substantially influencing the acrylic acid yield. The best yield was achieved with the nickel-manganese substituted catalyst at 30 bar reaction pressure with 15 bar partial oxygen pressure resulting in a total acrylic acid yield of 22 %. Therefore, 15 bar oxygen pressure was used for all following investigations.

Substrate to catalyst-ratio:

Due to the high dilution of the stock solution and the low density of acrolein approximately 38 mg substrate and 10 mg of catalyst (4:1 ratio) were used with all previous set of experiments in the parameter determination. The further investigation was performed with ratios of additional 3:1 , 2:1 , 1 :1 and 1 :2. The results are shown in figure 7. The values for conversion, selectivity and yields vary only by 6 % within the ratios for the set of catalyzed experiments. Compared to the blank experiments the acrylic acid yield is doubled. The results show the trend of decreasing selectivity with higher loads of catalysts above the ratio of 1 .5:1 . For further investigations the ratio is kept fixed at 1.5.1.

Substrate concentration:

Substrate concentrations were varied up to 10 vol.% of acrolein in water. The reactions with 10 vol.% were performed in glass inlets for the reactors. The resulting reaction solution was more viscous and deep blue compared to the orange-colored reaction solution at 1 vol.%. The results in figure 8 show, that the conversion of acrolein is increasing and the liquid overall yield decreases showing higher values of polymerization and gas phase products. Comparing the yield of acrylic acid, it is visible that a lower acrolein concentration is a better initial position for oxidation with less amounts of undesired by-products.

For the following experiments the fixed parameters can be summarized with a partial oxygen pressure of 15 bar, a total reaction pressure of 30 bar, a substrate:catalyst ratio of 1 .5: 1 , a freshly prepared substrate solution with a concentration of 1 vol.% acrolein in water and the use of the nickel-manganese polyoxometalate.

Design of experiments (DoE):

Due to the usage of a ten-fold set up the limit in the reaction planning was the variation of the temperature as a so-called hart-to-change factor (HTC). To ensure an optimal randomization of the planned experiments the software Design Expert 11 was used. The experimental planning was carried out using a split-plot design (due to the HTC factors) and a response surface optimization. The design type is best D-optimal and the design model is quadratic. For testing linearity two center points and for better knowledge seven additional model points and three additional groups were chosen. To be more independently from runaways one replicate is chosen for all points. As responses for analysis the conversion of acrolein, the yield of acrylic acid and the liquid phase selectivity for acrylic acid are chosen. The fixed parameters for the design are 15 bar partial oxygen pressure, 30 bar total reaction pressure, 1 vol.% acrolein in water as substrate and a substrate:catalyst ratio of 1.5:1. As a catalyst the nickel-manganese polyoxometalate (HiiPMnNiMoioC>4o) was used. The variables are the reaction time from 30 min to 240 min and the reaction temperature from 160 °C to 200 °C. The results are shown in table 3.

Table 3: Experimental plan with results shown for each run at reaction conditions. Reaction Temperature (T reaction), reaction time (Reaction), acrolein conversion (Xacroiein) , acrylic acid yield

(Y acrylic aci _d) and liquid phase selectivity (Sacryiic acid, liquid phase) are shown.

The highest acrylic acid yield was achieved at a reaction temperature of 180 °C with a reaction time of 135 min and 156 min. Here, the conversion of acrolein was ranging from 79 - 85 % with acrylic acid yields of 25 % and a selectivity ranging from 28 - 33 %. The achieved responses (acrolein conversion and acrylic acid yield and selectivity) were further analyzed using the DoE software. Hereby, the fraction design space (FDS) is analyzed first.

The FDS is the volume of the design space and shows whereas a prediction variance is less or equal to a specified value. The higher the calculated FDS value is, the better the data and the assumptions on it will be. For the conversion the Ad is 1 and the As is 0.55 resulting in an FDS value of 0.95, meaning 95 % of the design are precise enough to predict the mean within ± 1 %. Concluding this the experiments were successful regarding the usage of data for optimization in a DoE. In the following the responses will be shown separately starting with conversion.

1. Acrolein conversion:

The results for DoE analysis of acrolein conversion are shown in figure 9. It is clearly shown that there is a correlation of reaction time or reaction time and reaction temperature. The lowest conversion is observed at the lowest temperature with the shortest reaction time. The conversion is lowest at 160 °C after 30 min with 57 %. With rising reaction temperatures and reaction time, the conversion increases with a maximum at 200 °C and 180 min.

2. Yield of acrylic acid

The results for DoE analysis of acrylic acid yield are shown in figure 10. The yield of acrylic acid is lower for shorter reaction times and lower reaction temperatures. A plateau is from 90 min to 180 min and 175 °C to 195 °C.

3. Liquid phase selectivity

The results for DoE analysis of acrylic acid yield are shown in figure 11 . Comparing the results to the contour plot of the acrylic acid yield the same plateau is visible with a maximum from 90 min to 180 min and 175 °C to 195 °C for liquid phase selectivity.

4. Optimization

In the next step, an optimization is carried out numerically using Design Expert 11 . The criterium is the maximization of yield and selectivity of acrylic acid with highest importance. The program fits the mathematical derived 3-dimensional plot to its maxima. To see, what the model calculates outside the beforehand chosen ranges the limits were broadened from 10 - 300 minutes and 140 - 220 °C. The highest acrylic yield of 23 % ± 2.6 % and selectivity 30 % ± 1.9 % with a conversion of 85 % ± 4.7 % is predicted at a reaction temperature of 188 °C and a reaction time of 121 min with a desirability of 0.88. The desirability is an index of how easy the goals are achievable. The value of 0.876 is compared to 1 easy to reach. The 2-dimensional contour plots for the broadened range in the optimization are shown in figure 12.

Experimental confirmation of DoE experiments: The reaction time variation was performed at 188 °C reaction temperature. Hereby, samples were taken every 10 min. The conversion, selectivity and product distribution are shown in figure 12.

The conversion of acrolein is 47 % after 10 minutes of reaction time and increases fast with ongoing reaction time. The same is visible for the selectivity, it reaches a maximum of 31 % after 50 min and stays steady within the ongoing reaction time. The experiments performed at a reaction time of 121 min confirmed the DoE predictions resulting in an average yield of acrylic acid of 23 % with a liquid phase selectivity of 28 % and an acrolein conversion of 92 %

Further, reactions were performed using crude acrolein instead of pure acrolein. Although the crude acrolein included different impurities, the resulting acrylic acid yields were comparable to those achieved with the purified acrolein. Using crude acrolein as a product from the glycerol dehydration (step one of the inventive process) is essential regarding a green process route from glycerol to acrylic acid.

Claims

1 . A process for preparing acrylic acid from glycerol comprising i. dehydration of glycerol to acrolein, and ii. oxidation of acrolein to acrylic acid, wherein (i) and (ii) are performed in a liquid phase in the presence of a homogeneous catalyst and a solvent.

2. The process according to claim 1 , wherein the homogeneous catalyst is a polyoxometalate.

3. The process according to the preceding claim, wherein the dehydration of glycerol (i) is performed in the presence of a polyoxometalate having a Keggin-structure H_a[XMi2C>4o], wherein

X is a hetero atom (preferably Si or P),

M is one or more different metals, preferably one, two or three different metals, wherein at least one metal is a transition metal, preferably Mo, W, or V, and a is 3+n, wherein n is the charge of [XMi204o]ⁿ‘.

4. The process according to the preceding claim, wherein the polyoxometalate is H4SiWi204o, H3PW12O40 or H3PM012O40.

5. The process according to any one of the preceding claims, wherein the dehydration of glycerol (i) is performed in the presence of a glycerin ester as the solvent, preferably glycerol monostearate.

6. The process according to any one of the proceeding claims, wherein the dehydration of glycerol (i) is performed at 150 to 300 °C.

7. The process according to any of the preceding claims, comprising additionally a distillation of acrolein into a separate reaction vessel subsequent to the dehydration of glycerol (i).

8. The process according to any one of the preceding claims, comprising addition of an inhibitor of the polymerization of acrolein and/or acrylic acid subsequent to the dehydration of glycerol (i) and/or the oxidation of acrolein (ii).

9. The process according to the preceding claim, wherein the inhibitor of the polymerization of acrolein and/or acrylic acid is selected from phenol, thiazine, derivatives and/or mixtures thereof, preferably from hydroquinone, 4-methoxyphenol, phenothiazine and/or mixtures thereof.

10. The process according to claim 2, wherein the oxidation of acrolein (ii) is performed in the presence of a polyoxometalate having a Keggin-structure K_a[XMi2C>4o], a Lindqvist- structure K_a[MeOi9] or a Wells-Dawson structure K X2M18O62], wherein K is H, an alkali metal or ammonium,

X is a hetero atom (preferably Si or P),

M is one or more different metals, preferably one, two or three different metals, wherein at least one metal is a transition metal, preferably Mo, W, V, Ni, Nb, Mn, Co or Cu, and a is 3+n, wherein n is the charge of [XMi204o]ⁿ‘, [MeO-ig]"' or [X2MisO62]ⁿ'.

11 . The process according to the preceding claim, wherein the polyoxometalate is H15PC03M09O40, HnPMnNiMoio04o or KIO 2WI?CO062-

12. The process according to any one of the preceding claims, wherein the oxidation of acrolein (ii) is performed in the presence of oxygen.

13. The process according to any one of the preceding claims, wherein the oxidation of acrolein (ii) is performed at 100 to 200 °C and a pressure of 1 to 80 bar, preferably at 160 °C to 200 °C and a pressure of 30 to 50 bar.

14. The process according to any one of the preceding claims, wherein acrolein is in the form of an aqueous solution with a concentration of < 20%, preferably with a concentration of < 10%

15. A composition comprising or consisting of a reaction product prepared by the process according to any of the preceding claims.

16. The composition according to the preceding claim, wherein the reaction product comprises acrylic acid and one or more of formic acid, acetic acid, glycolic acid and/or glycerol.

17. The composition according to any one of claim wherein the reaction product comprises acrylic acid in an amount of > 7 mol% and formic acid in an amount of < 50 mol%.