Master Thesis Development of a Semi-continuous Set-up for Catalytic Hydrotreatment of Lignin

(1)

Master Thesis

Development of a Semi-continuous Set-up for Catalytic Hydrotreatment of Lignin

Author: A.J. Telgenhof (s1947389)

Supervisors: Prof. dr. ir. H.J. Heeres Prof. dr. F. Picchioni Daily supervisor: Ir. A. Kloekhorst

Institute for Technology, Engineering and Management Chemical Technology

August 2013

(2)

A.J. Telgenhof - Master thesis - August 2013 - ITEM - Chemical Technology Page 2 of 92

(3)

Abstract

Depletion of fossil fuels and the ongoing production of greenhouse gasses have led to an increased demand of renewable energy, transportation fuels and chemicals. Whereas energy can be produced by using wind and solar power, chemicals need a carbon containing source. Biomass is the only sustainable source of carbon available at the moment for producing chemicals. Lignocellulosic biomass is the woody part of biomass and is made out of three main components, being cellulose, hemicellulose and lignin. Lignocellulosic biomass does not compete with food and has enormous potential as a feedstock for second generation bio-fuels and chemicals. This study focuses on a specific part of lignocellulosic biomass, namely lignin. Lignin has an enormous potential for the production of fuels and bulk chemicals, due to its aromatic structure [1]. Aromatic structures are valuable, because of the high demand of phenols and benzene, toluene and xylene (BTX). This study focuses on the valorization of lignin with the use of heterogeneous catalysis and hydrogen under elevated conditions (100-160 bar and 400°C). The objective of this study was to develop an optimal semi-continuous set-up for catalytic hydrotreatment of lignin towards phenolic and aromatic products. A catalytic screening, pressure optimization and influence of reaction time were

performed. Five catalysts were found to be suitable for the catalytic hydrotreatment of lignin. Total aromatic and phenolic yield was in the following order Ru/TiO2 (12.9%) (ruthenium chloride made) >

Ru/TiO2 (12.0%) (ruthenium acetate made) > Cat. D (11.9%) > Ru/C (11.4%) > Pd/C (8.1%). Ruthenium on titania (ruthenium chloride made) therefore was the most suitable catalyst for maximizing the total phenolic and aromatic yield. Light oil yield is very important in semi-continuous catalytic hydrotreatment of lignin. Combining total phenolic and aromatic yield with the light oil yield, ruthenium on carbon was found to be the most suitable catalyst. At 130 bar the light oil yield for ruthenium on carbon was the highest, 15.7%. Therefore 130 bars was found to be the optimal pressure for the semi-continuous catalytic hydrotreatment of lignin with ruthenium on carbon.

Reaction time had large influence on the amount of light oil produced for a ruthenium on carbon catalyst, it increased over time (2.2% at 0.5 h to 28.9% at 8 h). Water amount also increased over time, where the oil yield decreased with about the same amount (75.0% at 0.5 h to 54.0% at 8 h).

Product elemental composition showed that the amount of oxygen was decreasing in the heavy oil and 0.22wt% at 8 h of reaction time. ALCELL lignin contains for 28.9wt% of oxygen. Total phenolic and aromatic yield increased with reaction time to a maximum at 4 h, from 5.3% at 0.5 h to 14.2% at 4 h.

(4)

1 Introduction

Depletion of fossil fuels and the ongoing production of greenhouse gasses has led to an increased demand of renewable energy, transportation fuels and chemicals. Whereas energy can be produced by using wind and solar power, chemicals need a carbon containing source. Biomass is the only sustainable source of carbon available at the moment for producing chemicals. Evidential

achievements have been made for example on the production of ethanol from starch and sugar rich components, which is called first generation bio-ethanol. However this process is competing with food and is therefore unwanted. Second generation bio-ethanol is made from lignocellulosic biomass [2,3]. Lignocellulosic biomass is made out of three main components, being cellulose, hemicellulose and lignin. Lignocellulosic biomass does not compete with food and has enormous potential as a feedstock for second generation bio-fuels and chemicals. This study focuses on a specific part of lignocellulosic biomass, namely lignin. Studies in the past have mainly focused on cellulose and hemicellulose valorization, where lignin has received lesser attention. In 2004 only the paper industry already produced 50 million tons of lignin out of paper waste streams. However the paper industry uses lignin as a low value fuel, because there is no demand for pure lignin at the moment. Only 2% of the available lignin is sold [1]. Still lignin has an enormous potential for the production of fuels and bulk chemicals, due to its aromatic structure [1]. Aromatic structures are valuable, because of the high demand of phenols and benzene, toluene and xylene (BTX). These chemicals are used for the production of numerous polymers, fuel additives, dyes, resins and pharmaceuticals [1]. This study focuses on the valorization of lignin with the use of heterogeneous catalysis and hydrogen under elevated conditions, which are 100-160 bar and 400°C.

In the first chapter of the thesis two literature reviews are presented, the first will describe the catalytic hydrotreatment of lignin or lignin derived compounds for selected catalysts. Most studies have been performed in a batch set-up, however this study will be done semi-continuous catalytic hydrotreatment of lignin. Therefore a second review is presented on semi-continuous catalytic hydrotreatment of lignin. With this work in mind the objectives and approach of this study will be presented. Experiments have been done to gain more insight in the use of semi-continuous heterogeneous catalytic hydrotreatment of lignin. First an overview will be presented about the structure of lignocellulosic biomass and more specific, lignin.

(7)

Literature review 1.1

1.1.1 Lignocellulosic biomass

Lignocellulosic biomass is made from carbondioxide and water using sunlight as the energy source, producing oxygen as a subproduct. This biochemical process is also known as photosynthesis.

Lignocellulosic biomass consists out of three different materials, being cellulose, hemicellulose and lignin. These materials all have different structures, as can be seen in Figure 1.

Figure 1: Structure of biomass [3]

Cellulose

Cellulose is a crystalline material with a wide, flat, 2-fold spiral structure, which makes up 40-80 weight percentage of the biomass. Cellulose consists of a linear polysaccharide with β-1,4 linkages of D-glucopyranose monomers, shown in Figure 2 [3].

Figure 2: Cellulose structure [3]

Chains of the cellulose are linear and flat and are strengthened by the hydrogen bonds.

Polymerization degree of cellulose in for example wood and cotton is approximately 10000 to 15000 glucopyranose monomer units [3]. Cellulose is broken down by partial acid hydrolysis into cellobiose (glucose dimer), cellotriose (glucose trimer) and cellotetrose (glucose tetramer). When complete acid hydrolysis takes place cellulose is completely broken down into glucose [3].

(8)

A.J. Telgenhof - Master thesis - August 2013 - ITEM - Chemical Technology Page 8 of 92 Hemicellulose

Hemicellulose is a sugar polymer that makes up 20-40 weight percentage of biomass [3]. The polymer is amorphous because of its branched nature. Hemicellulose can be distinguished from cellulose by the number of sugars. Where cellulose only has one sugar, hemicellulose has five different sugars. An example of hemicellulose is shown in Figure 3.

Figure 3: Galactoglucomannan, a structure of hemicellulose [4]

This complicated polysaccharide appears together with cellulose in the cell walls. The materials contains five-carbon sugars (usually xylose and arabinose) and six-carbon sugars (galactose, glucose and mannose). Xylan is the most occuring building block of hemicellulose (a xylose polymer linked at the 1 and 4 positions). Cellulose is hard to hydrolyze, where hemicellulose is easy to hydrolyze to its monomer compounds [3].

Lignin

Lignin is a polyaromatic material, which is 10-30 weight percentage of lignocellulosic biomass and 40% by energy [1]. In the next chapter lignin will be described further, because this study demands for a more detailed description of this interesting compound.

(9)

A.J. Telgenhof - Master thesis - August 2013 - ITEM - Chemical Technology Page 9 of 92 1.1.2 Lignin

Lignin structure

Lignin is an amorphous polymer, which exists out of three-dimensional methoxylated phenylpropane structures. In plants lignin fills up the gaps between the cellulose and hemicellulose and acts like a natural resin, which gives strength to the structure [1].

Figure 4: Lignin in plant cells [1]

In Figure 4 the place of lignin in the plant cells is shown. There has been done much research on the structure of lignin, but an exact composition is difficult and differs per plant species. It is believed that the biosynthesis of lignin takes places by polymerization of p-coumaryl, coniferyl and sinapyl alchohols, shown in Figure 5 [1].

Figure 5: Building blocks of lignin [1]

Amount, composition and molecular weight of lignin are different for each plant, but the amount will normally decrease in the following order, softwoods > hardwoods > grasses. An example of a lignin type in a softwood is shown in Figure 6 [1].

(10)

A.J. Telgenhof - Master thesis - August 2013 - ITEM - Chemical Technology Page 10 of 92 Figure 6: Softwood lignin [1]

In softwood lignin only p-coumaryl and coniferyl alcohol are presents, while in hardwood coniferyl and sinapyl alcohol are most prevalent in the structure. Softwood lignin contains about 90 % coniferyl alcohols, where in hardwood lignin coniferyl and sinapyl incorporate about the same amount. Typical linkage between the monomer units are shown in Table 1. The most common linkage in lignin is β-O-4, which contains about 40-45% in softwood and 60-62% in hardwood. This linkage is used as model compound in various studies for example by Jongerius et al. [5]. Coniferyl or sinapyl alcohols are connected by a Cϒ-OH group. 5-5 linkage is the second most appearing

compound in softwood lignin, for hardwood it depends on the type of wood, this linkage is also studied by Jongerius et al. [5]. Also carbon-carbon bonds are present, mainly as the 5-5 compounds [1].

(11)

A.J. Telgenhof - Master thesis - August 2013 - ITEM - Chemical Technology Page 11 of 92 Table 1: Linkages found in lignin [1]

Linkage β-O-4 5-5 β-5 Spirodienone

Abundange Softwood 40-45 19-27 9-12 0-2

per 100 C9-units Hardwood 60-62 0-9 3-11 0-5

Linkage 4-O-5 β-1 Dibenzodioxocin β-β

Abundange Softwood nd-7 1-9 0-7 2-6

per 100 C9-units Hardwood nd-9 1-7 0-2 3-12

1.1.3 Lignin treatment

After the biomass is harvested from the land, it has to be pre-treated, before it can be chemically handled. Biomass generally not only contains cellulose, hemicellulose and lignin, but also proteins, lipids, soil salts, water and ash. All these impurities have to be separated. In this work the focus will be on the treatment of lignin. There are several ways to extract lignin from biomass, which are described in section 1.1.3.1. A scheme of a bio-refinery for different lignin treatments is shown in Figure 7.

Figure 7: Bio-refinery [1]

(12)

A.J. Telgenhof - Master thesis - August 2013 - ITEM - Chemical Technology Page 12 of 92 After the biomass pre-treatment, the lignin can be catalytic treated, pyrolyzed or gasificated. The

focus of this study is on catalytic depolymerization, which is described in section 1.1.4. The goal of this catalytic depolymerization is to produce low molecular weight high value chemicals, for example phenolics. Phenolics can be used in many applications. In section 1.1.3.2 a depolymerization reaction of a typical lignin model compound (β-O-4) is shown, which produces numerous phenolics [5].

1.1.3.1 Lignin pre-treatment

According to recent work, there are four different separation methods available for lignin, being:

physical pre-treatment, solvent fractionation, chemical pre-treatment and biological pre-treatment.

[1] Various lignin sources can be used as a feedstock in a lignin bio-refinery. These different materials can arise from the paper industry (e.g., kraft and lignosulfunate lignin) or from other resources (e.g., organosolv lignin). Every method has it advantage and disadvantage, which will be described further for kraft, lignosulfonate and organosolv process. An overview of the process types with the molecular formula and weight is shown in Table 2.

Table 2: Overview of lignin processes [1]

Type Monomer molecular formula Monomer molecular weight

Kraft lignin C₉H_8.5O_2.1S_0.1(OCH₃)_0.8(CO₂H)_0.2 180

Technical kraft lignin C9H7.98O2.28S0.08(OCH3)0.77 176,52

Unreacted kraft lignin C₉H_8.97O_2.65S_0.08(OCH₃)_0.89 189,73

Lignosulfonate lignin (softwood) C9H8.5O2.5(OCH3)0.85(SO3H)0.4 215-254 Lignosulfonate lignin (hardwood) C₉H_7.5O_2.5(OCH₃)_0.39(SO₃H)_0.6 188

Organosolv lignin C9H8.53O2.45(OCH3)1.04 not defined

Pyrolysis C₈H_6.3-7.3O_0.6-1.4(OCH₃)_0.3-0.8(OH)_1-1.2 not defined

Steam explosion lignin C9H8.53O2.45(OCH3)1.04 ~188

Dilute acid lignin C₉H_8.53O_2.45(OCH₃)_1.04 ~188

Alkaline oxidation lignin C9H8.53O2.45(OCH3)1.04 ~188

Beech lignin C₉H_8.83O_2.37(OCH₃)_0.96 not defined

Kraft lignin process

Kraft lignin process is the most used in the industry, where under high pH, temperatures between 150-180°C, and large amounts of water with dissolved sodium hydroxide and sodium sulphide, lignin is separated [1]. This process is mainly used in the paper industry, where lignin is separated from the cellulose. While cellulose and hemicellulose are used for making paper, lignin is a waste stream and used as fuel to heat the processes.

Lignosulfonate process

Another process which is commonly used in the paper industry is the lignosulfonate process. Here lignin is separated form lignocellulosic biomass between pH 2 and 12, where calcium and magnesium sulfite is used. This process delivers a product, which is soluble in water [1]. Lignosulfonated supplies higher average molecular weights and monomer molecular weights than the kraft lignin.

(13)

A.J. Telgenhof - Master thesis - August 2013 - ITEM - Chemical Technology Page 13 of 92 Organosolv process

As described before the organosolv lignin differs from conventional methods used in the paper industry. The most well-known organosolv process is the ALCELL process, from which the product is used in this study. This process has been operated in the past by Repap Enterprises, where lignin is solved in ethanol or ethanol/water mixtures [1,6]. The organosolv process produces lignin, cellulose and hemicellulose streams, which is a great advantage, while the streams can be valorized

separately. Also the process is more environmental friendly than processes in the paper industry, because there are no sulfides used and reaction conditions are mild. Lignin from the organosolv process contains low values of sulphur, which is a great advantage. Sulphur is incorporated by other lignin pre-treatment methods and can cause deactivation of catalysts [1].

ALCELL lignin pre-treatment process

In this research ALCELL lignin, which is an organosolv lignin is used as feedstock and as used, this pre- treatment process will be described in detail. Kraft processes have been used for some time and have been proven economically feasible, but the environmental issues demand more sustainable

processes. The ALCELL organosolv process has been proven to be economically and more environmental friendly. Figure 8 shows the process scheme of the ALCELL process [7].

Figure 8: ALCELL process [7]

As described before in the organosolv process, individual biopolymers are separated using water or a water/ethanol mixture. The process is performed in counter current mode, where the plant pulp is mixed with the solvents, which is called cooking liquor. Temperature of the process lies between 180-210°C or 195°C and elevated pressure [6,7]. There are three separation steps. In the first separation step, hemicellulose, furfural, acetic acid and low molecular lignin is separated from the stream. The second and third step purifies the lignin to a higher quality. During the process the liquor is recovered and can be used again. In this process there are two streams of lignin, the final

concentrated black liquor and the low molecular weight lignin in the first stream. Lignin is separated from the concentrated black liquor and is sold as ALCELL lignin.

Alcell pulp is made from northern hardwoods and has a Canadian standard freeness of 400 ml, which is a quality indication [8]. Pulps from northern hardwoods contain about 5-15% of softwood fibers [7].

(14)

A.J. Telgenhof - Master thesis - August 2013 - ITEM - Chemical Technology Page 14 of 92 1.1.3.2 Depolymerization of lignin

Lignin forms a 3D network of different compounds. Many studies have been done on the catalytic hydrotreatment using model compounds, some are described in Table 4. Depolymerization of lignin can be described in more detail by looking at the reaction pathways of those model compounds.

Zakzeski et al. described model compound studies on β-O-4, carbon-carbon, β-5, α-O-4, α-O-5, p- coumaryl alcohol, coniferyl alcohol and sinapyl alcohol [1]. In this section depolymerization of compounds with the most common lignin linkage β-O-4 is described [5].

β-O-4

(1) + H2 →

4-(1-hydroxyethyl)-2- methoxyphenol

↓

Syringol

(2) + 2 H₂ → Methane + Water

↓ Guaiacol

(3) + H₂ → Methane

↓

Catechol Phenol

(4) + H₂ → + Water

Scheme 1: Depolymerization of β-O-4 linkage at 300°C, 50 bar H₂ for 4 h [5]

Table 3: Products of β-O-4 linkage at 300°C, 50 bar H2 for 4 h [5]

Conversion (%) 100

Mass balance (%) 33

Phenol (%) 9

0-Cresol (%) 2

p-Cresol (%) 2

Dimethylphenol (%) 5

Catechol (%) 1

Resorcinol (%) 2

Guaiacol (%) 6

Methoxyphenol (%) 2

Syringol (%) 4

(15)

A.J. Telgenhof - Master thesis - August 2013 - ITEM - Chemical Technology Page 15 of 92 Some typical reactions, which can occur are described in Scheme 1. This reaction pathway is

described by Jongerius et al. with the use of a sulfided cobalt molybdenum catalyst (CoMo) [5]. The first reaction describes the reaction of β-O-4 into syringol (4% conversion on β-O-4 intake), a well- known lignin model compound. Syringol is hydrodeoxygenated towards guaiacol (8% conversion on β-O-4 intake), which causes the release of methane and water. The release of methane is called demethylation [5]. Guaiacol reacts with hydrogen to catechol (1% conversion on β-O-4 intake), where just like in the second reaction, demethylation takes place. Finally the catechol converters to phenol (9% conversion on β-O-4 intake) by HDO. These reactions are a few of the many which take place in the depolymerization of lignin. All products which are made during this reaction are shown in Table 3. Reactions like the one presented in Scheme 1 will be dependent on reaction conditions and catalytic activy. Section 1.1.4 will describe the influence of catalysts on hydrotreatment of lignin in detail. Figure 9 shows some pathways, which are preferable for the production of chemicals.

Figure 9: New technologies in lignin treatment [1]

To describe the depolymerization of lignin there are five typical reactions that probably take place, which are:

Hydrogenation reactions: addition of hydrogen to C=C bonds.

Hydrodeoxygenation reactions: breaking up of C-O bonds where water is formed, which are shown in reaction 2 and 4.

Demethylation reactions: breaking up of C-O bonds where methane is formed, which are shown in reaction 2 and 3.

Decarboxylation reactions: breaking up of bonds where oxygen is removed in the form of carbondioxide or carbonmonoxide.

Thermal cracking reactions: breaking up of lignin towards lower molecular weight components by thermal energy.

These reactions will take place depending on the type of catalyst used, which will be described in the following section.

(16)

A.J. Telgenhof - Master thesis - August 2013 - ITEM - Chemical Technology Page 16 of 92 1.1.4 Catalytic treatment of lignin

There are many different catalytic methods for depolymerization of lignin, catalytic cracking or hydrolysis, catalytic reduction and catalytic oxidation reactions. Lignin oxidation, a process developed in the paper industry will give more complex compounds, where more functionalities are present.

Hydrolysis of lignin takes place with the use of water and catalytic cracking is a process where zeolites are used to depolymerize lignin. Reduction of lignin as described before, produces platform chemicals like phenolics and BTX, which have high demands in the chemical industry. Lignin

reduction has a high potential in the chemical industry and therefore this study will focus on this process and will be described further [1].

Lignin reduction/catalytic hydrotreatment

Where most publications on catalytic hydrotreatment of biomass are focused on the upgrading of bio-oil to turn them into transportation fuels, lignin reduction focuses on conversion towards platform chemicals [1]. Overall for all biomass derived products decreasing the oxygen level is very important, because this increases the energy value and decreases the viscosity of the oil [9]. Catalytic hydrotreatment lowers the oxygen content and reduces molecular weight. Three types of catalysis are possible for lignin reduction, being heterogeneous catalysis, homogeneous catalysis and electrocatalysis. Homogeneous catalysts are dissolved in a medium and electrocatalysis is a technique where the catalysis takes place in an electrochemical cell [1]. Heterogeneous catalysis is used in this study, because of the ease of separation of the products and the catalyst.

1.1.4.1 Heterogeneous catalysis

Many research on heterogeneous catalytic hydrotreatment has been done in the past, which are described in numerous reviews [1,9-12]. One of the first and widely accepted reviews is written by Furimsky. This review describes the catalytic hydrotreatment of biomass. The review is written on the basis of model compound studies [10]. Elliot describes the developments in hydrotreatment of bio-oils for many processes used at the moment [9]. Elliot has also published many papers on the hydrotreatment of model compounds, from which some are used in this study [13-17]. An essential review for this study is written by Zakzeski et al. on the catalytic valorization of lignin. This review gives an insight in most of the research, which has been done on the catalytic hydrotreatment of lignin [1]. Mortensen et al. wrote a review on the catalytic upgrading of bio-oil, which focused on the hydrotreatment reactions [11]. Gasser et al. produced a review with some information about multi- catalytic valorization of lignin [12].

Heterogeneous catalysts review

This study focuses on the use of heterogeneous noble metals, iron and nickel-copper catalysts for the catalytic hydrotreatment of lignin. Iron and nickel-copper metals are selected for their low

production costs. Noble metals are chosen, because of their ability to produce high oil yields and deoxygenated oils. In Table 4 a selection of the most relevant literature for this study is presented and description follows. Papers are described on oil yield and products composition as well as reaction conditions.

One of the first articles published about the catalytic hydrotreatment of lignin is by Harris et al., which describes the reaction of lignin with hydrogen over a copper-chromium oxide at moderate temperature (260°C) and high pressure (220 bar). They found an oil yield of 70% on lignin intake, where the main compounds were methanol, 4-n-propylcyclohexanol and 4-n-propylcyclohexanediol [18]. Brewer et al. also used a copper-chromium oxide catalysts under nearly identical conditions and a shorter reaction time, but found lower oil yields. So increasing the reaction time has probably a positive influence on the oil yield. In this paper 4-n-propylcyclohexanol was also found in the

(17)

A.J. Telgenhof - Master thesis - August 2013 - ITEM - Chemical Technology Page 17 of 92 products of the reaction [19]. Noble catalysts were studied by Pepper et al., being rhodium and

palladium on carbon and alumina supports. Palladium on carbon gave the best result on oil yield (72% on lignin intake) at moderate conditions (195°C and 34 bar). Rhodium on carbon gave an yield of 52% on lignin intake and rhodium on alumina had the lowest yield (48% on lignin intake). Carbon shows to be a more active support than alumina with a rhodium catalyst. Yields reported were higher than previous research under the same reaction time, which indicates that noble catalysts have higher activities than copper-chromium oxide catalysts. Products which were found by Pepper et al.

were dihydroconiferyl alcohol and 4-n-propylguaiacol [20,21]. A paper, which describes the use of iron oxide catalysts is written by Koyama, where several monomers were hydrocracked using sulfided iron oxide catalysts. It was found that iron oxide supported on alumina had slightly higher conversions than unsupported iron oxide, because it has a higher bond cleavage activity [22].

Shaptai et al. did research on lignin model compounds with various noble catalysts on a gamma alumina support . Catalysts were ruthenium, iridium, rhenium, palladium, iron, rhodium, platinum and nickel. Results of the screening are shown in Figure 10 and 11.

Figure 10: Activity in C-O hydrogenolysis Figure 11: Activity in ring hydrogenation (350°C and 137 bar) [23] (350°C and 137 bar) [23]

Activity in C-O hydrogenolysis of the catalysts was reported to decrease as follows, palladium >

rhodium > rhenium > iridium > ruthenium > platinum > iron > nickel. Ring hydrogenation reactions were also reported and were in the following order, palladium > rhenium > iridium > rhodium >

platinum > nickel > ruthenium > iron [23].

Lignin slurry oil was used by Meier et al. for catalytic hydropyrolysis experiments, where palladium on active carbon and iron oxide catalysts were tested. Oil yields were about 81% and 17% on lignin intake for palladium on active carbon support and iron oxide respectively. Palladium on carbon gave the lowest chary yield, which was about 1% on lignin intake where for iron oxide the amount of char was significant (about 50% on lignin intake) [24].

(18)

Page 18 of 92 Table 4: Literature on heterogeneous catalytic hydrotreatment of lignin or lignin model compounds

Process conditions

Entry Process Catalyst Support T (°C) P (bar) t (min) Lignin (model) compound Major products Conversion/Yield (%) Solvent Ref

1 Batch Cu-CrO None 260 220 1080 Lignin

Methanol,

4-n-propylcyclohexanol,

4-n-propylcyclohexanediol 70 Methanol/glycol [18]

2 Batch Cu-CrO None 250 200 300 Hydrol lignin

3-cyclohexyl-1-propanol, 4-n-propylcyclohexanol,

3-(4-hydroxycyclohexyl)-1-propanol 12 Methanol/ethanol/water [19]

3 Batch Rh Carbon 195 34 300 Spruce wood meal

dihydroconiferyl alcohol,

4-n-propylguaiacol 52 Dioxane/water [20]

4 Batch Rh Al2O3 195 34 300 Spruce wood meal

5 Batch Pd Carbon 195 34 300 Spruce wood meal

6 Batch Rh Carbon 195 34 300 Aspen wood meal Unknown a Dioxane/water [21]

7 Batch Fe2O3 None^b 450 98 50 Dimeric species Benzenes, monophenols, dimers 3-100 None [22]

8 Batch Fe2O3 Al2O3b 450 98 50 Dimeric species Benzenes, monophenols, dimers 12-100 None [22]

9 Batch M^c Al2O3b,d 350 137 a

Diphenyl ether and naphthalene mixture

Phenol, benzene, cyclohexane,

tetralin, decalin a n-pentadecane [23]

10 Batch Pd

Activated

charcoal 380 100 15 Organocell lignin Phenolics, cresols, guaiacols, catechols 81 None [24]

11 Batch Fe2O3 None 380 100 15 Organocell lignin Phenolics, cresols, guaiacols, catechols 17 None [24]

12 Batch Pd Carbon 250 50 30 Phenol Cyclohexanol, cyclohexane 100 Phosphoric acid/water [25]

13 Batch Pd Carbon 250 50 30 4-n-propylguaiacol Cycloalkanes, cycloalcohols, methanol 100 Phosphoric acid/water [25]

14 Batch Pd Carbon 250 50 30 4-allylguaiacol Cycloalkanes, cycloalcohols, methanol 99 Phosphoric acid/water [25]

15 Batch Pd Carbon 250 50 30 4-acetonylguaiacol Cycloalkanes, cycloalcohols, methanol 100 Phosphoric acid/water [25]

16 Batch Pd Carbon 250 50 30 4-allylsyringol Cycloalkanes, cycloalcohols, methanol 92 Phosphoric acid/water [25]

18 Continue Rh SiO2 300 10 a Anisole Phenol, phenol derivatives 30 None [26]

19 Continue Rh ZrO2 300 10 a Anisole Phenol, phenol derivatives 91 None [26]

20 Continue Rh CeO2 300 10 a Anisole Phenol, phenol derivatives 95 None [26]

21 Continue Ni-Cu Al2O3 300 10 a Anisole Phenol, phenol derivatives 99 None [26]

22 Continue Ni-Cu ZrO2 300 10 a Anisole Phenol, phenol derivatives 60 None [26]

23 Continue Ni-Cu CeO2 300 10 a Anisole Phenol, phenol derivatives 100 None [26]

24 Batch Pd Carbon 200-300 69 240 Guaiacol

Volatile hydrocarbons, cyclohexanediol,

2-methoxycyclohexanol 66 Acetic acid/water [14]

25 Batch Ru Carbon 200-300 69 240 Guaiacol 2-methoxycyclohexanol, cyclohexanol 100 Acetic acid/water [14]

a Not specified, ^b Sulfided, ^c M = Ru, Cu, Ir, Re, Pd, Fe, Rh or Pt, ^d ϒ-Al2O3, ^e Made from RuCl3, ^f Made from Ru(acac)3, ^g Ru(NO3)3.

(19)

Page 19 of 92

Process conditions

Entry Process Catalyst Support T (°C) P (bar) t (min) Lignin (model) compound Major products Conversion/Yield (%) Solvents Ref

26 Batch Ru Carbon 200 40 240 Lignin Monomers, dimers 15 Water [27]

27 Batch Pd Carbon 200 40 240 Lignin Monomers, dimers 34 Water [27]

28 Batch Rh Carbon 200 40 240 Lignin Monomers, dimers 27 Water [27]

29 Batch Pt Carbon 200 40 240 Lignin Monomers, dimers 42 Water [27]

30 Batch Pd Carbon 250 40 30-120 Monomers, dimers Alkanes, methanol 95-100 Water [27]

31

Semi-

continue Ru Carbon 350 100 120 Pyrolytic lignin oil

Cycloalkanes, alkyl-substitute cyclohexanols,

cyclohexane and linear alkanes a Dodecane [28]

32 Batch Ru Carbon 170 40 240 Eugenol Phenolics 91 Isopropanol [29]

33 Batch Pd Carbon 170 40 480 Eugenol Phenolics 91 Isopropanol [29]

34 Batch Ru SBA-15 170 40 240 Eugenol Phenolics 100 Isopropanol [29]

35 Batch Pd SBA-15 170 40 480 Eugenol Phenolics 73 Isopropanol [29]

36 Batch FeOx-ZrO2-Al2O3 None 300 10 120 Lignin Phenol, cresol, methoxy phenol, alkyl phenol 88 Butanol/water [30]

37 Batch Pt Carbon 250 20 180 Corn stalk lignin

4-ethylphenol,

4-ethylguaiacol 72 Ethanol/water [31]

38 Batch Ru Carbon 200-275 20-60 60-120

Corn stalk lignin, bamboo lignin

4-ethylphenol,

39 Batch Pd Carbon 250 20 180 Corn stalk lignin

4-ethylphenol,

40 Batch Ru Carbon 200 30 240 Kraft lignin

Propylguaiacol, ethylguaiacol, methylguaiacol,

guaiacol a Ethanol/water [33]

41 Batch Pd Carbon 200 30 240 Kraft lignin

42 Batch Pt Al2O3 200 30 240 Kraft lignin

43 Batch Ru Carbon 250 100 260 Phenol Cyclohexanol, cyclohexane 88 Dodecane [34]

44 Batch Ru^e Carbon^e 250 100 260 Phenol Cyclohexanol, cyclohexane 90-100 Dodecane [34]

45 Batch Ru^f Carbon^f 250 100 260 Phenol Cyclohexanol, cyclohexane 90-100 Dodecane [34]

46 Batch Ru^g Carbon^g 250 100 260 Phenol Cyclohexanol, cyclohexane 90-95 Dodecane [34]

47

Semi-

continue Ru Carbon 350, 400 80-160 240 ALCELL lignin

Phenolics, alkylbenzenes, naphtalenes, alkanes,

methane, carbonmonoxide, carbondioxide 38-55 None [35]

a Not specified, ^b Sulfided, ^c M = Ru, Cu, Ir, Re, Pd, Fe, Rh or Pt, ^d ϒ-Al2O3, ^e Made from RuCl3, ^f Made from Ru(acac)3, ^g Ru(NO3)3.

(20)

A.J. Telgenhof - Master thesis - August 2013 - ITEM - Chemical Technology Page 20 of 92 Recently Zhao et al. published a paper on the catalytic conversion of lignin model compounds to

alkanes with a palladium on carbon catalyst. These experiments showed that palladium is a very effective catalysts for the catalytic hydrotreatment of phenol, 4-n-propylguaiacol, 4-allylguaiacol, 4- acetonylguaiacol and 4-allylsyringol towards cycloalkanes, cycloalcohols and methanol [25].

A continuous process of lignin model compounds is presented by Yakolev et al., who describe the hydrotreatment of anisole with rhodium and nickel-copper on some interesting supports. Ceria was found to be the most effective support for hydrotreatment for both catalysts, 95% and 100% yield on anisole intake for rhodium and nickel-copper respectively. Alumina also showed to be a very good support for the nickel-copper catalyst (99% yield on anisole intake). Main products of the reactions were phenol and its derivatives [26].

Catalytic hydroprocessing of guaiacol was done by Elliott et al. with the use of palladium and ruthenium on carbon. Results of two comparable experiments are shown in Figure 12 and 13 at 69 bar and 250°C.

Figure 12: Ruthenium on carbon on hydro- Figure 13: Palladium on carbon on hydro-

Processing of guaiacol (250°C and 69 bar for 4 h) [14] Processing of guaiacol (250°C and 69 bar for 4 h) [14]

Ruthenium showed to be a more active catalyst for hydrotreatment of guaiacol than palladium on a carbon support, 100% conversion against 66%. Palladium does not catalyze aqueous-phase reactions, so it can be used at higher temperatures than ruthenium, which gasificates the guaiacol at high temperatures. Palladium produces large amounts of methanol, where ruthenium procuses large amounts of methane [14].

Yan et al. produced a paper with a two-step degradation of wood lignin over noble catalysts. Four noble catalysts were used to for hydrotreatment of wood lignin towards monomers and dimers, being ruthenium, palladium, rhodium and platinum on a carbon support on modest temperature and pressure. The total yield decreased in the following order, platinum > palladium > rhodium >

ruthenium, where platinum produced a yield of 42% on lignin intake of monomers and dimers.

Monomers which were produced are guaiacylpropane, guaiacolpropanol, syringolpropanol and syringolpropane, dimers are not specified. Platinum was selected for the second step in the two-step process, where it gave excellent results in hydrotreatment of the monomers and dimers [27]. De Wilt et al. produced a paper on the hydrotreatment of pyrolytic ALCELL lignin with the use of a ruthenium on carbon catalyst in a semi-continuous process. They concluded that ruthenium on carbon gave high yields towards low molecular phenolics. The reaction produces lots of cycloalkanes, alkyl-substitute cyclohexanols, cyclohexane and linear alkanes, where aromatic structures are wanted [28]. A recent paper on the hydrotreatment of phenolic compounds was written by Guo et al., where palladium and ruthenium catalysts on carbon and SBA-15 (mesoporous silica) supports was investigated. Ruthenium on a SBA-15 support was found to produce the highest conversion, this was even the case at lower reaction temperature or shorter reaction times. At 100°C the ruthenium on SBA-15 gave an 87%

conversion and at 80°C the conversion dropped to 29% [29]. Another recent publication was written about the production of phenols from lignin with the use of an iron oxide-zirconia-alumina catalyst. It was found that a temperature of 300°C, a pressure of 10 bars and 120 minutes reaction time gave an

(21)

A.J. Telgenhof - Master thesis - August 2013 - ITEM - Chemical Technology Page 21 of 92 oil yield of 88% on lignin intake and a phenolic yield of 3% on lignin intake [30]. A paper by Ye et al.

describes the hydrotreatment of two types of lignin (corn stalk and bamboo) with noble catalysts (palladium, ruthenium and platinum on carbon). Two components were selected for analyzing purposes (4-ethylphenol and 4-ethylguaiacol). Ruthenium gave the best results in selected product yield measured [31]. Higher hydrogen pressure results in a higher yield of 4-ethylphenol. Sufficient amounts of hydrogen can prevent repolymerization reactions, which was already concluded by Meier et al., who hydrocracked lignin at 375-425°C [32]. Zakzeski et al. did experiments on lignin

valorization for the production of aromatic chemicals and hydrogen. Platinum on alumina and

ruthenium and palladium on carbon were used at 30 bar, 200°C and 4 h of reaction time. Platinum on alumina gave the best result in the formation of useful products, while palladium on carbon gave the second best results and ruthenium on carbon the lowest yields. In the catalytic lignin valorization process adding of ethanol will hinder the repolymerization of lignin, because it interacts with the disrupted linkages. In the absence of ethanol this polymerization results in the formation of highly recalcitrant solids (char) [33].

Wildschut et al. did experiments on the hydrotreatment of phenol with different ruthenium carbon catalysts. Three different self-made carbon supports were tested on ruthenium loading and

compared with the ruthenium on carbon commercial catalyst. Results of the experiments are shown in Figure 14.

Figure 14: Ruthenium experiments on hydrotreatment of phenol (250°C and 100 bar for 4.3 h) [34]

Two products were measured, being cyclohexanol and cyclohexane. High catalyst loadings gave higher yields towards cyclohexane for all catalysts, but only the RuCl3 based catalyst had no mass balance losses at high loadings, therefore being the superior catalyst [34]. This study is an extension of the work of Huisman. His work is described in section 1.1.4.2 [35].

Conventional catalysts

Well known catalysts for the hydrotreatment of biomass are cobalt and nickel promoted

molybdenum (CoMo and NiMo), which have proven to be effective in the oil industry for HDS and HDN treatment. These catalysts have the property to be effective for coal and oil with oxygen contents up to 10% and for biomass even up to 50%. [10] Some of these catalysts are described in Table 5 [24,32,36-38].

(22)

A.J. Telgenhof - Master thesis - August 2013 - ITEM - Chemical Technology Page 22 of 92 1.1.4.2 Reaction conditions

Literature on reaction conditions is also important and is shown in Table 5. Wildschut et al. did experiments on catalytic hydrotreatment of pyrolysis oil. Two different conditions were tested with ruthenium on alumina, titania and carbon, palladium on carbon, platinum on carbon, NiMo on alumina and CoMo on alumina. Experiments were done at 250°C and 100 bar for all catalysts, results of product distribution is shown in Figure 15.

Figure 15: Products for hydrotreatment of Figure 16: Gas distribution for hydrotreatment of pyrolysis oil (250°C and 100 bar for 4 h) [36] pyrolysis oil (250°C and 100 bar for 4 h) [36]

Closure of the mass balances was difficult for alumina supports. Ruthenium on titania and carbon and platinum on carbon gave the best results. Figure 16 shows the gas production, mainly carbondioxide was released. The formation of carbondioxide is probably caused by the

decarboxylation of organic acids. Deep hydrotreatment (350°C and 200 bar) was also tested with the same catalysts, results of these experiments are shown in Figure 17 and 18.

Figure 17: Products for hydrotreatment of Figure 18: Gas distribution for hydrotreatment of pyrolysis oil (350°C and 200 bar for 4 h) [36] pyrolysis oil (350°C and 200 bar for 4 h) [36]

(23)

A.J. Telgenhof - Master thesis - August 2013 - ITEM - Chemical Technology Page 23 of 92 Table 5: Catalytic hydrotreatment literature on reaction conditions

Process conditions

Entry Process Catalyst Support T (°C) P (bar) t (min) Feedstock

H2 flow

(ml/min) Major products Conversion/Yield (%) Solvent Ref

1 Batch

Ru, Pd, Pt, NiMo, CoMo

Al2O3a

, TiO2,

carbon 250, 350 100, 200 240 Fast pyrolysis oil -

Phenolics, alkylbenzenes, naphtalenes, alkanes, methane, carbonmonoxide,

carbondioxide 20-65 None [36]

2 Batch Ru Carbon 350 200 60-600 Fast pyrolysis oil -

Aldehydes, ketones, lignin monomers,

hydrocarbons, carbondioxide, methane 40-50 None [39]

3 Semi-continue

NiMo^c,

Cr2O3d Al2SiO5c

,

Al2O3d 400 50-140 150-480 Organocell lignin 500

Phenol, 3,5-dimethylphenol, 4-

propylphenol, 3,4-dimethylphenol 30-80 None [37]

4 Batch

NiMo^c,e, Cr2O3d

Al2SiO5c

, Al2O3d

,

zeolite^e 395-430 90-100 20-60 Several Lignins -

Alkyl benzenes, phenols, polycyclic

aromatics 49-71 None [38]

5 Semi-continue

NiMo, CoMo,

zeolite None^f 375-450 75-180 120 Organocell lignin b

Phenol, cresols, alkylphenols, xylenols,

guaiacols 28-83

Hot separator

top, lignin oil [32]

6 Batch

Pd, Fe2O3, NiMo^c,g,h, Raney-Ni

Activated charcoal, Al2SiO5c,g

,

zeolite^h 380 100 15 Several Lignins - Phenolics, cresols, guaiacols, catechols 17-81 None [24]

7 Semi-continue Ru Carbon 350, 400 80-160 240 ALCELL lignin 10-200

carbondioxide 38-55 None [35]

8 Batch Ru Carbon 350 100 120

Pyrolytic lignin

oil -

Cycloalkanes, alkyl-substitute cyclohexanols, cyclohexane and linear

alkanes b Dodecane [28]

9 Batch Ruⁱ Carbonⁱ 250, 350 100, 200 260 Pyrolysis oil, -

carbondioxide 40-60 Dodecane [34]

a ϒ-Al2O3,^b Not specified, ^c M8-81 catalyst, ^d M9-10 catalyst, ^e M8-85 catalyst, ^f Some catalysts were sulfided, ^g M8-82 catalyst, ^h M8-86 catalyst, ⁱ Commercial and made from RuCl3, Ru(acac)3 and Ru(NO3)3.

(24)

A.J. Telgenhof - Master thesis - August 2013 - ITEM - Chemical Technology Page 24 of 92 In these experiments ruthenium on titania and palladium on carbon gave the best results regarding oil yields and ruthenium and palladium on carbon gave the best results on oxygen removal. Palladium and platinum on carbon were the most active catalysts, because all the hydrogen was consumed. Support influence on oil yield for ruthenium decreased as followed, TiO2 > C > Al2O3 and for oxygen removal, C >

Al2O3 > TiO2 [36].

Additionally Wildschut et al. did experiments with a ruthenium on carbon catalysts for hydrotreatment of pyrolysis oil on. All experiments were done at 350°C and 200 bars and the reaction time was varied.

Results of the products distribution versus time are shown in Figure 19.

Figure 19: Products for hydrotreatment Figure 20: Gas distribution for hydrotreatment of pyrolysis oil (350°C and 200 bar) [39] of pyrolysis oil (350°C and 200 bar) [39]

The amount of solids decrease, with increasing of the time. According to the paper this is within the experimental error. The amount of water increased, this is typical for dehydration and HDO reactions.

Gas distribution of the product is also measured, this is shown in Figure 20. Formation of carbondioxide and methane during this experiment is typical for ruthenium on carbon experiments on pyrolysis oil. [39]

A semi-continues process is described by Meier et al. for the hydropyrolysis of organocell lignin. The effect of reaction time and pressure were tested using a 1:1 mixture of two catalysts, NiMo on aluminosilica and chromium oxide on alumina. The experimental set-up used is shown in Figure 21.

Figure 21: Experimental semi-continuous set-up of Meier et al. [37]

(25)

A.J. Telgenhof - Master thesis - August 2013 - ITEM - Chemical Technology Page 25 of 92 Results of the experiments are shown in Figure 22 and 23.

Figure 22: Reaction time for hydrotreatment Figure 23: Pressure on hydrotreatment of lignin of lignin (400°C and 120 bar) [37] (400°C for 6 h) [37]

Reaction time experiments were done at a temperature of 400°C and a pressure of 120 bar. At 360 minutes the oil yield was the highest and the char yield the lowest of all tests. The amount of light oil increased with time and was 50% on lignin intake at 360 and 480 minutes. Influence of pressure showed that increasing the pressure gives better results in oil yield and result in low char production [37].

Oasmaa and Meier described the hydrotreatment of different types of lignin with NiMo on aluminosilica, NiMo on zeolite and chromium oxide on alumina catalysts. NiMo catalysts were tested in their oxide and sulfide forms. Most experiments were done with mixed catalysts (1:1) in a batch reactor. Oil yield produced was in the following order, organocell lignin > pine kraft lignin > birch kraft lignin. The

detectable oils ranged from 13.9 to 38.4 weight% of lignin [38]. Another article published by Meier et al.

describes the hydrocracking of organocell lignin using several oils as solvent. A significant decrease in solid formation is observed in the presence of the NiMo catalyst and a significant amount of hydrogen (180 bar). The solids amount decreases from 23.8 to 0.6% and the oil yield on lignin intake increases from 53.7 to 81.9% [32].

Meier wrote an article about the influence of time, pressure and reaction time with a NiMo catalyst [24,37]. In Figure 24 the results of the experiments are shown.

Figure 24: Influence of temperature (100 bar for 0.25 h)(l), pressure (420°C and 0.25 h)(m) and reaction time (420°C and 100 bar)(r) for catalytic hydrotreatment of lignin [24]

Four temperatures were tested in finding the most efficient condition, the performance of the catalysts was the best at 400°C. Increasing amount of char was due to the shortage of hydrogen.

It was found that complete depolymerization will not take place at 350°C, because high molecular weight lignin particles were present in the acetone extract after the reaction took place. Only guaiacol was detected in the heavy oil at 350°C. The amount of monomeric fraction gave an optimum at 100 bars, but

(26)

A.J. Telgenhof - Master thesis - August 2013 - ITEM - Chemical Technology Page 26 of 92 the oil yield is the highest at 120 bars. Oil yields are independent of reaction time, but decreased at 120 minutes. The amount of char stayed roughly the same [24].

Previous work on the use of ruthenium on carbon has been done by Huisman. This work contained pressure optimization of the semi-continuous catalytic hydrotreatment of lignin, which is presented in Figure 25 and 26 [35].

Figure 25: Optimization of oil yield on semi-continue catalytic hydrotreatment of lignin (Ru/C and 400°C for 4 h) [35]

Figure 26: Oil and char yield on semi-continue catalytic hydrotreatment of lignin (Ru/C and 400°C for 4 h) [35]

The optimal pressure for oil and char yield lies between 100 and 130 bar. Light oil yield is the highest at 130 bar [rep3].

De Wilt et al. wrote an article about lignin valorization, which is described in section 1.1.4.1 [28].

0 10 20 30 40 50 60 70 80 90 100

80 100 130 160

Yield (weight/weight lignin) [%]

Pressure (bar)

Water Light oil Heavy oil Char

0 10 20 30 40 50 60 70 80 90 100

80 90 100 110 120 130 140 150 160

Yield (weight/weight lignin) [%]

Pressure (bar)

Light oil Total oil Char

(27)

A.J. Telgenhof - Master thesis - August 2013 - ITEM - Chemical Technology Page 27 of 92 In addition to the work of Wildschut et al shown in section 1.1.4.1, hydrotreatment of fast pyrolysis oil is described. Figure 27 shows the difference in product distribution on the influence of different ruthenium on carbon synthesis methods and loadings.

Figure 27: Product distribution for catalytic Figure 28: Gas distribution for catalytic hydrotreatment of pyrolysis oil hydrotreatment of pyrolysis oil (350°C and 200 bar for 4.3 h) [34] (350°C and 200 bar for 4.3 h) [34]

The commercial ruthenium on carbon catalyst shows the highest oil yield, but also gives the highest char yield and deoxygenation. The amount of methane formed is related to the amount of ruthenium loaded on the support, which can be seen in Figure 28 [34].

Master Thesis Development of a Semi-continuous Set-up for Catalytic Hydrotreatment of Lignin

Master Thesis

Development of a Semi-continuous Set-up for Catalytic Hydrotreatment of Lignin

Abstract

Table of contents

1 Introduction

Literature review 1.1