Direct Liquefaction of Biomass - Wiley Online Library

Chem. Eng. Technol. 2008, 31, No. 5, 667–677

667

Frank Behrendt1 York Neubauer1

Review

Michael Oevermann1 Birgit Wilmes1 Nico Zobel1

Direct Liquefaction of Biomass

1

Berlin Institute of Technology, Department of Energy Engineering, Berlin, Germany.

Reserves of fossil primary energy carriers are limited. Consequently liquid secondary energy carriers especially for mobile applications made from fossil reserves will not carry on forever but need to be replaced in a not-to-far future. Two substitution strategies are currently under investigation – the use of oil from plant seeds either directly or after chemical modification (biodiesel) or the gasification of complete plants, use of the product gases (mainly CO and H2) in a Fischer-Tropsch process with subsequent refining. A third possible pathway would be the so-called direct liquefaction, i.e., the conversion of complete plants into liquid fuels without gasification. This process is discussed and various technical implementations are critically evaluated in the present paper. Keywords: Biomass, Liquefaction Received: February 4, 2008; accepted: March 25, 2008 DOI: 10.1002/ceat.200800077

1

Introduction

Reserves of fossil primary energy carriers are limited. Consequently also liquid secondary energy carriers especially for mobile applications made from fossil reserves will not carry on forever but need to be replaced in the not to far future. Two substitution strategies are currently under investigation: – 1st generation bio fuels: Use of oil from plant seeds either directly or after chemical modification (biodiesel) as well as fermentation products from glucose (ethanol), or – 2nd generation bio fuels: gasification of complete plants, use of the product gases (mainly CO and hydrogen) in a Fischer-Tropsch process with subsequent refining towards high-quality fuels A third possible pathway would be the so-called direct liquefaction, i.e., the direct conversion of complete plants into liquid fuels without the gasification step. This one-step process is discussed and evaluated in the present paper. The current status of this process in literature is summed up. Various technical implementations are discussed in some detail. Direct liquefaction of coal (coal to liquid – CtL) under increased pressure in the presence of hydrogen and a catalyst is an about 100 years old process investigated in some detail in the first half of the last century mainly by Bergius and Pier [1–4]. A detailed review of the development of CtL has been given by Kaneko [8]. Further details about the process can be found in [5, 6, 8, 9].

– Correspondence: Prof. Dr. F. Behrendt ([email protected]), Berlin Institute of Technology, Department of Energy Engineering, Chair for Energy Process Engineering and Conversion Technologies for Renewable Energies, Fasenenstraße 89, D-10623 Berlin, Germany.

© 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

The indispensable need for hydrogen present in the process not only for coal but also for biomass becomes obvious when the conversion is considered in detail. Dry wood can be described by the sum formula CH1,4O0,7, liquid fuels by CH2. Consequently the chemical basis of direct liquefaction as well as the technical challenge can be described by the reaction equation (1): CH1,4O0,7 → CH2

(1)

Hydrogen has a double function: Adaptation of the C/H ratio of biomass to the one of liquid fuels and simultanous removal of oxygen from the raw material. The chemical reactions involved in this process build a complex reaction network. As early as during the first oil crisis in the first half of the seventies of the last century direct liquefaction of biomass came into focus as one potential path away from crude oil as the basis for mobility. The misfortune of these early approaches was not only based on the – compared with the prices of crude oil – too high process prices. Additionally, the lack of basic scientific understanding of the process led to the failure of early pretty large demonstration plants built during that period. In the mean time a large number of investigations worldwide resulted in a much better understanding of the process. In Germany some of the current technical implementations of direct liquefaction of biomass apply conditions very similar to those used with coal as a raw material while others believe being able to run the process without additional hydrogen and/or without increased pressure. A number of the current process implementations in Germany are based on adventurous assumptions with respect to the chemical reactions involved. In some cases the reaction equation given above is taken without sufficient reflection as the basis for the process design. An additional and very critical

http://www.cet-journal.com

668

shortcoming of many of the processes is the lack of understanding of the chemical steps involved. Moreover when aiming at producing liquid fuels for transportation purposes the product has to comply with the respective national and international standards for these fuels, i.e., with EN 228 for Otto fuels and EN 590 for Diesel fuels.

2 2.1

Table 1. Selected parameters from the European standards for Otto (EN 228, top) and Diesel (EN 590, bottom) fuels. Parameters of Otto fuel (EN 228)

OK 2000 98/70/EG

OK 2005 98/70/EG

Maximum concentration of benzene (vol.-%)

1,0

1,0

Maximum concentration of aromatic compounds (vol.-%)

42

35

Maximum concentration of sulfur (ppm)

150

50

Basic Aspects of Direct Liquefaction of Biomass

Maximum concentration of olefines (vol.-%)

18

18

Maximum concentration of oxygen (wt%)

2,7

2,7

Structure of Biomass

Parameters of Diesel fuel (EN 590)

DK 2000 98/70/EG

DK 2005 98/70/EG

Cetane number

51

51

Maximum concentration of polyaromatic compounds (wt%)

11

11

Maximum concentration of sulfur (ppm)

350

50

Minimum density (g/L)

820

820

Maximum density (g/L)

840

840

The main constituents of biomass are the biopolymers cellulose, hemicellulose and lignin. Furthermore, a number of inorganic compounds are present representing the major source of ashes when the biomass is combusted. For thermochemical conversion processes leading to small molecules not only these principal constituents are important but also the general structure of the plant. This general structure will influence the early process steps like the milling of the plant as well as chemical or thermal decomposition of the raw material. The wall of lignocellulotic plant cells typically consists of 30–35 wt% cellulose, 15–35 wt% hemicellulose and 20–35 wt% lignin. Softwood has a relatively higher content of lignin compared with hardwood while hardwood contains more hemicellulose [10]. Fresh biomass contains up to 60–70 % water, after drying this value is reduced to about 10–15 %.

2.2


F. Behrendt et al.

Table 2. Operational parameters and products for direct liquefaction, flash pyrolysis, pyrolysis and gasification Process

Solvent

Pressure

Temperature

Product

Direct liquefaction

Yes

< 1 – 240 bar

150 – 420°C

Organic liquid

Flash pyrolysis

No

< 1 – 5 bar

< 500°C

Organic liquid

Pyrolysis/Gasification

No

< 1 – 20 bar

700 – 900°C

Solids and gas

Requirements of Fuels

Acidic Hydrolysis Basic Hydrolysis 䊊 Neutral Hydrolysis – Organic Medium – Thermal decomposition under reducing atmosphere. 䊊䊊

As pointed out a key criterion for a viable direct liquefication process is the quality of the products. Here the relevant European standards for liquid fuels have to be obeyed with respect to each and every point. In Tab. 1 some of these parameters are summarized.

2.3

Reaction Conditions, Differentiation from Pyrolysis and Gasification

For direct liquefaction of biomass as well as pyrolysis and gasification of biomass a large variety of reactor types, process conditions and media (solvents) have been applied. Tab. 2 gives an overview of important parameters and also shows the differences of direct liquefaction as compared to pyrolysis and gasification. Direct liquefaction itself can be further classified into different reaction paths and reaction conditions. Chornet suggested the following classification [11]: – Solvolysis – Aqueous Medium


3

Conceivable Reaction Pathways for Direct Liquefaction of Biomass

The conversion of wooden biomass into liquid hydrocarbons comprises the following steps [11]: 1. preparation of feedstock 2. slurrying the feedstock within a liquid carrier 3. heating the slurry to reaction conditions 4. addition of reducing gas (e.g. H2 or H2/CO) at elevated pressure 5. main reaction 6. product separation, and 7. solid-liquid separation and recovery of solvent Usually catalysts of various functions are added to the slurry. These engineering process steps of liquefaction correspond to the following chemical steps:



I. II.

solvolysis of the biomass depolymerization of the main components cellulose, hemicellulose, and lignin III. chemical and thermal decomposition of monomeres and smaller molecules leading to new molecular rearrangements through bond ruptures, dehydration, and decarboxylation IV. degradation of oxygen containing functional groups in the presence of hydrogen. The de-polymerization of cellulose and hemicellulose in aqueous media can be regarded as relatively well known. As cellulose and hemicellulose are polysacharides (pentoses and hexoses) a huge knowledgebase from the sugar and pulp and paper industries exists. Depending on substrate and temperature (hemicellulose 120–180 °C, cellulose > 240 °C) the de-polymerization in aqueous media under the addition of acid is followed by rupture of ring bonds or re-arrangement of the formed monosaccharides [12]. Characteristic for the de-polymerization of cellulose and hemicellulose is the huge number of different degradation products with a multitude of oxygen containing functional groups. From these functional groups only a few are able for further direct reduction of oxygen by cleavage of carbon dioxide. Therefore, the presence of hydrogen is essential for the degradation of functional groups and to avoid re-polymerization. The de-polymerization of lignin in a solvent leads to numerous different substituted phenols. As in the above mentioned case for cellulose and hemicellulose the absence of hydrogen quickly leads to re-polymerization of the products.

4

Experimentally Established Reaction Pathways

Only a few publications on direct liquefaction of biomass focus on the analysis of intermediate and final products and the development of reaction pathways. In the review articles by Chornet et al. [11] and Bouvier et al. [13] information on the de-polymerization of the main macromolecules lignin, cellulose, and hemicellulose depending on used media is given. Further reaction steps can usually be found in articles dealing with specific liquefaction processes under simplified conditions [12, 14–20]. They are usually limited to one of the following processes: – liquefaction of cellobiose and lignin model components in phenol [15, 19–21] – liquefaction of cellulose in undercritical water [12, 14, 22] – liquefaction of “Kraft” lignin in tetralin [23–25] and – hydrodeoxygenation of lignin model components with hydrogen [26] The investigations of reaction mechanisms and kinetic data in the cited literature mostly deal with model substances in homogeneous phases. However, it seems to be questionable if the results of these simplified setups can be transferred and used for real biomass [12] as biomass contains different types of salt causing a considerably modification of the reaction pathways. In addition, during the liquefaction of biomass organic molecules with numerous functional groups are present


Liquefaction

669

opening a wide range of different reaction pathways making a systematic investigation of reaction paths somewhat impossible. Some general information on the reactions during the liquefaction of lignin, cellulose, and hemicellulose can be found in [10, 11, 13]. In these articles liquefaction is classified based on the carrier medium being acidic, alkaline, or neutral. For the liquefaction process of biomass the following targets can be identified [11, 13]: – solvolysis in micelle-like substructures – rupture of wood macro molecules – stabilization of reactive species – reduction of oxygen concentration, and – increasing the H/C ratio De-polymerization can be achieved by sequential dissolving of macro molecules through utilization of their physical and chemical properties (e.g. paper production), or integral dissolving, i.e. parallel dissolving of all types of macromolecules. Following some general comments on reaction pathways for biomass de-polymerization by integral dissolving are given while neglecting the interactions between the macromolecules and intermediate products. In general, hydrolysis of carbohydrates and polysaccarides in an alkaline medium is slower than in an acid one. Hydrolysis in acid media. Cleavage speed of polysaccharides depends on the applied acid medium and the properties of the slurry. The glycosidic bonds of cellulose are ruptured at elevated temperatures of 200–220°C. De-polymerization leads to the formation of glucose units [13] degrading further into hydroxymethyl-furfural [11]. De-polymerization of hemicellulose into mono saccharides starts at temperatures lower than 120 °C [13] with a selectivity depending on the solvent used and the pH value. In a subsequent hydrolysis step the formed monosacharides are converted into furfural and acid derivatives. For lignin, a-OH and a-O groups are converted into benzylium ions [11]. At 160–180°C in weak acid media a-ethers groups are ruptured [11, 13]. In a subsequent step the benzylium ions are either sulfonated or condensated. Hydrolysis in alkaline media. Cellulose and hemicellulose are converted at temperatures > 140 °C into organic acids by rupture of glycosidic bonds [13]. The OH– groups of lignin favor the de-lignification through rupture of ester bonds and a-ether bonds lead to stilbene structures whereas methoxyl groups are left over in form of methanol in the presence of hydroxyl anion. In addition, condensation reactions lead to diaryl methane structures and – by oxidation – to chromophoric groups. Solvolysis in organic media. The reactions in an organic medium strongly depend on the interactions between solvent and substrate. As a rule of thumb the solvent should be chosen such that it strongly reacts with cellulose. If possible the solvent should be a product of the liquefaction itself, e.g. phenol and its derivatives, simple alcohol or polyalcohol. The homolytic rupture of C-O and C-C bonds lead to short-chained hydrocarbons. Besides these general comments one can find some more detailed reaction pathways for lignin and cellulose model substances in the literature to be summarized for lignin and cellulose below.


670

F. Behrendt et al.

Lignin. In a series of publications Lin et al. [19–21, 27, 28] investigated the liquefaction of biomass with phenol as a solvent and presented some complex reaction pathways. Besides cellobiose they investigated lignin model substances in some detail. They observed a broad range of intermediates and final products during liquefaction, e.g. 28 and 8 final products for lignin and cellobiose, respectively. In a less detailed study Connors et al. [23] investigated reaction pathways for lignin with H2 donators and tetralin as a solvent. For the model substance ethyl-guaiacol they identified two (methyl-O and the cyclic-O scission) and for the model substance dehydrodihydrodiisoeugenol three (1-a-, a-b-, b-5-scission) reaction paths. However, they concluded that the investigation of single reaction steps for lignin is not feasible. By investigating the hydrodeoxygenation of a lignin model substance under liquefaction conditions Petrocelli et al. [26] observed that the oxygen in lignin prohibits the thermal conversion of lignin into shorter chained products leading to substantial amounts of re-polymerized long chained components. They concluded that the product yield of benzene and cyclohexan compared with phenol can be improved substantially by reducing the oxygen content of lignin during liquefaction. In a recent publication by Fang et el. [15] the decomposition of lignin in supercritical water/phenol solutions under high temperature conditions was investigated (see Fig. 1). They proposed reaction paths for lignin in supercritical water consisting of four different phases: – oil phase (phenolics, PAHs, heavy HCs), – aqueous phase (acids, aldehydes, alcohols, catechol, phenols), – gas phase (CO2, CO, H2, C1–C4 HCs), and – a solid residue phase.


Fang et al. concluded that lignin can be completely dissolved and undergoes homogeneous hydrolysis and pyrolysis preventing further re-polymerization. Cellulose/glucose. Reaction pathways for cellulose in undercritical water in the presence of soda (Na2CO3) can be found in [12, 14, 18, 22]. They are depicted in Fig. 2. The reaction steps for the conversion of glucose in aqueous media are relatively well known and depicted in Fig. 3. The liquid products are organic acids, aldehydes, and alcohols, all carrying a substantial amount of oxygen. Fang et al. observed in [14] that without a catalyser these liquid components will lead to a high fraction of a so called glucose-coke, whereas in the presence of soda they will convert into oily products, and in the presence of Ni into mainly gaseous products. In a recent work by Huber et al. [17] a process for the liquefaction of carbohydrates from biomass for the production of liquid alkanes (C7–C15) is presented. The main reactions are dehydration and hydrogenation leading to a successive hydrophobing of the organic reactants. In addition to these reactions alkaline catalytic aldol reactions lead to the formation of large organic components which can be further converted into liquid alkanes. The work of Huber is of particular interest as it is – to our knowledge – the only reaction pathway leading direcly to liquid alkanes as final products. Most experimental investigations on liquefaction of biomass deal with the investigation of a single and often simplified model substance. For these model substances intermediate and final products can be determined with some detail and corresponding reaction pathways have been presented in the literature. However, the presented results do not allow a direct generalization and transferability to the liquefaction of real biomass. If hydrogen for the saturation of intermediate products is used in combination with further process and conditioning steps one is able to produce alkanes of different size distributions. This is an important precondition for the production of liquid fuels conforming to usual standards.

5

Figure 1. Reaction path of lignin in supercritical water [15].


Parameters Influencing the Yield of Direct Liquefaction

The gaseous, liquid, and solid yield of direct liquefaction of biomass depends on numerous parameters. These parameters can be classified into chemical and physical parameters. Their impact will be summarized in this section. But first, it should be emphasized that a meaningful comparison of experimental results between different research groups concerning the yields of direct liquefaction is usually not possible due to the following reasons:



Liquefaction

671

– the definitions of liquid and solid yield vary – the system pressure during liquefaction is often not documented, and – the product treatment before analysis differs significantly between different research groups Due to these difficulties quantitative comparisons between different direct liquefaction experiments are simply not meaningful. Therefore, the impact of chemical and physical parameters on the liquefaction yield can only be reviewed in qualitative terms.

5.1

Impact of Chemical Parameters

5.1.1 Lignin Content of the Biomass Figure 2. Reaction paths of cellulose in aqueous media [15].

Demirbas [29] concluded that the liquid yield decreases with increasing lignin content. This was not confirmed by Minowa et al. In their study [30] no significant relation between lignin content and liquid yield was found.

5.1.2 Solvent

Figure 3. Reaction paths for glucose in aqueous media [18].


Many investigations have been conducted on the influence of the solvent on the liquefaction yield. One of the most comprehensive studies with respect to this aspect was conducted by Mun et al. [31]. They found that the lowest solid residue content is achieved by use of simple alcohols (methanol, ethanol, propanol, butanol). Higher alcohols as well as organic acids result in much higher solid residue contents. However, the disadvantage of the simple alcohols is their relatively low evaporation temperature. They basically evaporate before the biomass gets liquefied. In an other study by Krzan et al. [32] it was demonstrated that even for chemically very similar solvents like propylene glycol, ethylene glycol, and diethylene glycol the liquid yields differ significantly (between 16 and 32 wt%).


672


F. Behrendt et al.

5.1.3 Catalyst Heterogeneous catalysts that are commonly used for hydrogenation (such as Raney-Nickel) do not exhibit catalytic activity towards biomass liquefaction [33, 34]. Homogeneous catalysts, however, are significantly more active than heterogeneous catalysts. Several species have been investigated as homogeneous catalysts: organic and inorganic acids [35], NaOH [36, 37], salts [37], and Ni-Oloat [34]. From these studies the following conclusions can be drawn: (1) using organic acids as liquefaction catalysts lead to a lower solid residue content than using inorganic acids, (2) salts (phosphates, sulfates, chlorides, acetates, carbonates) have lower catalytic activity towards liquefaction than NaOH at typical liquefaction temperatures.

the other way round. This interpretation is supported by the gas yields that where measured in dependence of the temperature during liquefaction.

5.2.2 Pressure According to Le Chatelier’s principle one would expect that the higher the system pressure during liquefaction, the less liquid components are gasified. This correlation was also found in experiments as shown in Fig. 5.

5.1.4 Atmosphere In three different studies [24, 33, 34] it was shown that the atmosphere is much less important than the solvent as far as the reduction ability is concerned. But it should be noted that the atmosphere has a strong influence on the oxygen content in the product oil.

5.2

Impact of Physical Parameters

5.2.1 Temperature Many research groups investigated the influence of the temperature on the oil-yield. The results of five of them are depicted in Fig. 4. Obviously the oil-yield reaches a maximum in an intermediate temperature range.

Figure 5. Pressure dependence of the oil yield as reported in different publications.

5.2.3 Mass Ratio of Solvent to Biomass (S/B) In three studies [38–40] it was found that the higher S/B ratio, the lower the solid residue content after liquefaction. This observation seems reasonable since with increasing solvent content more monomers can be solved. However, Maldas et al. [41] observed an opposite trend. This contradiction cannot be rationalized. Boocock et al. [42] examined the impact of the S/B ratio on the oil yield. They observed a vulcano-type behavior with a maximum oil yield in a S/B ratio range of 3 to 5. It is not clear, why the oil yield decreases at higher S/B ratios.

5.2.4 Concentration of the Homogeneous Catalyst

Figure 4. Temperature dependence of the oil-yield as reported by different research groups.

These results can be interpreted as follows: The higher the temperature, the easier the defragmentation of the polymers into a liquid oil-rich phase. A further increase of the temperature leads to enhanced decomposition of these fragments into gaseous species. Below a critical temperature the former process dominates the latter. Above this critical temperature it is


The solid residue content depending on the catalyst’s concentration (mass of catalyst per mass of solvent) was measured in [35, 43, 44]. Ogi et al. [45] investigated how the oil-yield depends on the catalyst concentration. The qualitative trends of all these experiments indicate that up to a critical value the oil yield increases (and the solid residue content decreases) with catalyst concentration. A further increase of the catalyst concentration above this critical value leads to lower oil yield (higher solid residue content). It is possible that the catalyst not only promotes de-polymerization but also re-polymerization. The latter process may dominate the former at high catalyst concentrations. But this assumption has to be confirmed by experiment.



5.2.5 Residence Time Similar to the temperature and the catalyst concentration dependence the influence of the residence time on oil yield obeys a vulcano-type behavior. Many research groups have investigated this correlation by experiment. One representative example for those findings is depicted in Fig. 6, where the results reported by Yilgin et al. [46] are illustrated.

Liquefaction

673

Obviously, the oxygen content in the product oil is significantly lower than in the original biomass. However, the oxygen content is still way too high to use the product oil as a gasoline or Diesel fuel substitute.

5.4 The BLTF Project The most ambitious international research project in the field of direct liquefaction, called “Biomass Liquefaction Test Facility”, was finished in 1985 [48]. Six laboratories conducted liquefaction experiments using exactly the same biomass with different experimental setups and reaction conditions. The product mixtures were examined in one laboratory for the sake of better comparability. The oil yields were in the range of 38 to 53 wt% (mass of oil to initial mass of biomass). The oxygen content in the oil ranged from 9 to 25 wt%.

6

Direct Liquefaction Processes

6.1

Introduction and Overview

There is a critical residence time for a maximum oil yield. The residence time is larger then the liquid components are further decomposed into gaseous components and re-polymerized to solid residue, respectively.

This section contains an overview of historic and ongoing approaches to the liquefaction process. A lot of knowledge in this research area was collected and based on this pilot plants are designed, installed and commissioned. In the following sections the processes listed in Tab. 3 are presented in more detail including reaction conditions and findings as far as published. All specifications concerning products, quality and amounts are based on the statements of the process owner.

5.3

6.2

Figure 6. Dependence of the liquefaction yield on residence time; data from [47].

Elemental Composition of the Product Oil

Several research groups have determined the elemental composition of the product oil. In Fig. 7 these results are summarized and compared with the values for wood and gasoline.

Historic Processes of Direct Liquefaction

In the context of the crude oil crisis the research activities in the field of direct liquefaction were intensified. Within this scope the PERC process (Pittsburgh Energy Research Center, Pittsburgh, USA) and the LBL process (Lawrence Berkeley Laboratory, Berkeley, USA) were developed. Both processes worked with a gas mixture consisting of CO2 and H2 generated by gasification of biomass. Economic reasons led to the abandonment of both processes despite intensive development efforts. High oxygen amounts in the biomass as well as the production of reaction water during hydrogenation turned out to be major drawbacks.

6.2.1 Pittsburgh Energy Research Center (PERC)

Figure 7. Elemental composition of the liquefaction product oil in terms of mass ratios as reported by different research groups compared to wood and gasoline.


This process works with wood chips and prepared product oil. The resulting mixture is pumped through a tube reactor with a residence time of 10 to 30 min at temperatures of 330 to 370 °C and a pressure of 200 bar. The oil yield amounts to 45 to 55 % of the employed dry matter of organic material. The recycle oil serves as a hydrogen supplier.


674


F. Behrendt et al.

Table 3. Overview of direct liquefaction processes of biomass and plastic wastes. Raw Material Developer/Supplier Process Name Liquefaction Liquefaction H2-atmosphere Catalyst of the Process Temperature Pressure [°C] [bar] Biomass

Pittsburg Energy Research Center (USA)

PERC-Process 330–370

200

Yes

Yes

Lawrence Berkeley LBL-Process Laboratory (USA)

330–360

170–240

Yes

Yes

Shell Research Institute (NL)

HTU-Process

265–350

180

No

No

BFH (GER)

BFH-process

380

100

Yes

No

HAW (GER)

DoS-Process

350–500

80

Yes

No

Müller & Bothur (GER)

B/M-process Mueborit

< 220

6

No

No

400

1

No

Yes

No

Yes

Giessen University LTC-process of Applied Sciences Plastic waste Ozmoenergy / Ozmoenergy Ozmotech Pty Ltd. (AUS) Gossler Envitech GmbH (GER)

Gossler

Clyvia Technology Clyvia CL500 GmbH (GER)

390–420

350–400

1

No

Yes

380–420

0.2–0.3

No

No

abling a fast heating of the aqueous solvent and a fast cooling of the reaction mixture. In 1984 a semicontinuous process for liquefaction of lignocellulosic materials was developed. A process based on catalytic hydrogenolysis using hydrogen, catalysts and oil showed best yields and promising carbon and energy balances. Initially hydrogen as reducing agent and oil recycled from the hot gas separator is dosed to the first reaction autoclave. The resulting conversion products are firstly fed to the hot gas separator (residence time: 15 min) and then charged into the cooling system to expand. Under these conditions and in the presence of palladium as catalyst wood flour could be hydrogenated. 100 % of wood results into 36 % oil as liquid tar, 50 % of carrier oil, which is reintroduced to the process, 5 % coke, and 25 % aqueous phase. A further optimization of the BFH process was stopped due to the better prospects using the flash pyrolysis.

6.2.2 Lawrence Berkeley Laboratory (LBL)

6.3

The LBL process starts with the hydrolysis of biomass with sulphuric acid which is followed by a neutralisation with sodium carbonate. The mixture is homogenized in a refiner and afterwards pumped through a tube reactor where it is liquefied at pressures of 100 to 240 bar and temperatures between 330 and 360 °C. Alkali carbonate is added to produce hydrogen in presence of water and CO2 at high pressures (water gas shift reaction) which is necessary to saturate free radicals. The product of the direct liquefaction is a liquid material similar to bitumen (high viscosity, 1.1–1.2 kg/m3, composition: 15–19 % oxygen, 6.8–8 % hydrogen and 74–78 % carbon). The upper heating value amounts to 34 MJ/kg.

6.3.1 Hydrolytic Process – HTU-Process (Shell)

6.2.3 Direct Catalytic High-Pressure Liquefaction – BFH Process In the 1980s the BFH (Bundesforschungsanstalt für Forst- und Holzwirtschaft, Germany) carried out an intensive research on the conversion of lignocellulosic raw and residual material into liquid products using thermal pressure treatment. Variation of reaction time, temperature, pressure as well as solvents, catalysts and reducing agents in the reaction system were investigated. The reactor system consists of 3 linked autoclaves en-


Current Processes of Direct Liquefaction

The investigations and development of fundamentals started in 1982 in the Shell Research Laboratory in Amsterdam (NL) by Goudriaan. After a short break in 1994 to 1997 an economic evaluation of the process was carried out. In 1997 to 2000 a project on process development followed. For 2008 / 2009 the starting up of a demonstration plant is planned with an annual throughput of 25.000 t/a. In the Hydrothermal Upgrading process a number of different biomasses (also with high moisture content) can be liquefied under high pressure [49–51]. The biomass is suspended and pumped into the reactor using a high-pressure pump. At temperatures of 300 to 350°C, pressures between 120 and 180 bar and a residence time of 5 to 20 min. a biocrude is produced. The oxygen in the biomass is removed by water and CO2. The product consists of 45 % biocrude (wt% of input material, dry and without ash), 25 % gas (> 90 % CO2), 20 % H2O and 10 % solved organic materials (e.g. acetic acid, ethanol). Biocrude is a heavy organic fluid that becomes solid at 80 °C. The heating value is 30–35 MJ/kg, the H/C ratio is 1:1 and the oxygen content is between 10 to 18 %. The thermal efficiency for one variant of this process amounts to 74.9 % (theoretically a maximum of 78.6 % could be reached) [49].



After that, the product is separated into a light and a heavy fraction which can be used in different energetic applications. Another option is the production of an oil fraction by an upgrading step using the HDO (hydro-deoxygenation) process being similar to the first stage hydrocracking processes HDS and HDN.

6.3.2 High-Pressure Hydrogenation Process/DoS Process The DoS process which was developed by the HAW (Hochschule für Angewandte Wissenschaften Hamburg, Germany) is a direct one-step liquefaction process for lignocellulosic biomass (e.g. wood, straw). It works under a pressure of about 80 bar and at temperatures between 350 and 500 °C. It is a bottom phase crack process based on a fast pyrolysis followed by the solvolysis into product oil. It is claimed that the required hydrogen is also generated from the biomass. The conversion of hackled and dried biomass is carried out in a bottom phase reactor under high-pressure using hydrogen to produce oil, water, coal and gas. Vaporous product oil is separated from water over the gas / vapor phase and is then, after condensation, gas separation and expansion, fed to the cascade distillation for fractionation. The solids leave the reactor together with the liquid phase and are thereafter separated from the oil being recycled. The thermal efficiency of the whole system is around 70 % based on the heating value of the charge [52]. A DoS pilot plant is at present under construction.

6.3.3 B/M Process – Mueborit A patented pulping process for biomass to produce liquid hydrocarbons using a catalyst was developed by Umwelttechnik Stefan Bothur in 1999 (Germany). This process is based on solvolysis. Using a discontinuous tank reactor at 6 bar and 200 °C lignocellulosic biomass are completely “solved” in a melt of potassium carbonate hydrate (30 % water). This “solving” process consists of a mixture of chemical and physical processes (decomposition of molecules, disproportioning, salification of the functional groups etc.). The solution mainly consists of ions of organic salts (e.g. carboxylates, alkoxides). Organic compounds are separated using a column (overhead product). Further separation techniques are under investigation. The product is a dark brown mixture consisting of different compound classes. The heating value of the products depends on the target yield. For yields up to 40 % a heating value of 35–37 MJ/kg was measured. The combustion of solid and highly viscous residues provides process heat.

6.3.4 Low Temperature Conversion of Organic Residual The low temperature conversion process (LTC) has been developed at the Giessen-Friedberg University of Applied Science by Stadlbauer [53]. The process operates under oxygen exclusion


Liquefaction

675

at atmospheric pressure and temperatures between 350–400 °C in the presence of a catalyst. The low temperature conversion process has been applied to the direct liquefaction of sewage sludge, meat and bone meal, tar sands, animal fat and fat residues, and plastic material. The reactor concept of the LTC utilizes a heterogeneous catalytic reaction between the organic gases from the biomass and the catalyst. The reactor was designed as a continuously mixed solid bed reactor with a spray condenser. The condensate is separated in a three-pass centrifuge into water, oil, and dust material. By NMR and infrared spectroscopy techniques the liquid oil products of the LTC at 400 °C have been identified as aliphatic hydrocarbons with a fraction of aromatic hydrocarbons of at most 5 %. The product oils are claimed to have comparable physical and chemical properties as Diesel fuel (heating value, oxygen content, emissions, etc.).

6.3.5 Thermofuel/Ozmoenergy A direct liquefaction process for the conversion of plastic waste and heavy oils into so-called “thermofuel” was patented by Ozmoenergy/Ozmotech Pty Ltd. (Australia) [www.ozmoenergy.com/technology/]. The plant has a capacity of 10–40 t/d. The following types of plastics can be used as raw materials: PE, PP, PS and mixtures of them. The company claims that roughly 1 L oil can be produced from 1 kg plastic waste. The company states that the “thermofuel” can be utilized in internal combustion engines and gas turbines. However, it is certainly not a Diesel fuel or gasoline.

6.3.6 Gossler Envitec Another process for the utilization of plastic waste (PE, PP, PS, EPS, PET) was developed by the German company Gossler Envitec GmbH. The liquefaction is conducted by use of a heterogeneous catalyst at ambient pressure. The product oil is probably similar to the “Thermofuel” of Ozmoenergy. Two pilot plants are currently running: the first one with a capacity of 3000 t/y in Korea, the second one with 3500 t/y in Germany.

6.3.7 Clyvia Technologies Heavy oils and plastic waste (PE, PP, PS) can also be converted by the CL500 process designed by Clyvia Technology GmbH (Wegberg-Wildenrath, Germany). After pre-heating and liquefaction of the raw material at 150–250 °C it is fed to a CSTR operating at 380–420 °C and ambient pressure. In the reactor the plastic waste gets depolymerized and evaporated. Subsequently the reaction mixture is separated in a distillation column into a bituminous and a light oil fraction. A pilot plant (4000 t/y) is running since mid 2006 in Wegberg-Wildenrath (Germany).


676

7


F. Behrendt et al.

Conclusions

The current activities with respect to direct liquefaction of biomass aiming at the production of liquid hydrocarbons mainly for application in transportation take place in more or less uncoordinated way. They are, generally, characterized by a low level of scientific understanding of the chemical and physical processes involved. On the basis of the limited data supplied by the process developers it is clear that simple one-step processes without the addition of hydrogen are not promising and should not be pursued in the future. An over-simplified description of the process (“CH1,4O0,7 (biomass) – CO2 = CH2 (fuel)”) may be tempting but ignores completely the structural aspects of the biomass and the vast amount of very reactive oxygenated intermediates of it decomposition to be dealt with. Hydrogen (under elevated pressure and in the presence of catalysts) does play an indispensable key role here in the decomposition of these intermediates towards finally water and CO2. This two-step approach may prove technically feasible. Besides this technical aspect also an economical feasibility must be given. An important point when comparing today’s fuel production from crude oil with, e.g., direct liquefaction of biomass is that the whole process chain from the raw biomass to the final product conforming to the standards has to be taken into account. In such analysis some costs are easily “forgotten”, e.g., the expenses for the supply of hydrogen. In conclusion the direct liquefaction of biomass compared with its elder relative – the liquefaction of coal – is far away from a technical and economical feasibility. The core problem is the presence of large amounts of oxygen to be removed before a useful fuel conforming to standards will result. Quite often successful steps or parts of the full process are viewed as a sign for the solution of the overall problem. To avoid major disappointment in the scientific community but also in politics and the public a careful analysis of claims of success is needed.

[6]

[7]

[8]

[9]

[10]

[11]

[12] [13] [14] [15]

[16] [17] [18]

Acknowledgement This review is based on a study prepared for the German Bundesanstalt für Landwirtschaft und Ernährung. The financial support through the contract 114-50-10-0337/05-B is gratefully acknowledged.

References [1] M. P. E. Berthelot, Les carbures d’hydrogene III, 1st ed., Gauthier-Villars, Paris 1901. [2] M. Höring, L. Raichle, Chem. Ing. Tech. 1966, 3, 205. [3] W. Krönig,(1950), Katalytische Druckhydrierung. Das I. G. Verfahren von Matthias Pier, 1st ed., Springer, Berlin, Göttingen, Heidelberg 1950. [4] M. Pier, Österreichische Chemiker Zeitung 1939, 2, 33. [5] Energy Research at DOE: Was It Worth It? Energy Efficiency and Fossil Energy Research 1978 to 2000, (Edited by US De-


[19] [20] [21] [22]

[23] [24] [25] [26] [27] [28] [29] [30] [31] [32]

partment of Energy), USA, 2001. http://darwin.nap.edu/ books/0309074487/html Coal Liquefaction - Technology Status Report 010 (Edited by The Department of Trade and Industry), UK, 1999. http:// www.berr.gov.uk/files/file18326.pdf T. Kaneko, F. Derbyshire, E. Makino, D. Gray, M. Tamura, Coal Liquefaction, in Ullmann’s Encyclopedia of Industrial Chemistry, Wiley-VCH, Weinheim 2002. H. Schmieder, E. Henrich, E. Dinjus, Wasserstoffgewinnung durch Wasserspaltung mit Biomasse und Kohle, Wissenschaftlicher Bericht FZKA 6556, Forschungszentrum Karlsruhe 2000. http://bibliothek.fzk.de/zb/berichte/FZKA6556.pdf F. P. Burke, S. D. Brandes, D. C. McCoy, R. A. Winschel, D. Gray, G. Tomlinson, Summary Report of the DOE Direct Liquefaction Process Development Campaign of the late Twentieth Century: Topical Report, DOE Contract DE-AC2294PC93054, 2001. http://www.osti.gov/bridge/product.biblio.jsp?osti_id=794281 D. Fengel, D. Grosser, Holz, Morphologie und Eigenschaften, 4th ed., in Ullmanns Encyklopädie der technischen Chemie, Vol. 12, Verlag Chemie, Weinheim - Deerfield Beach – Basel 1984. E. Chornet, R. P. Overend, Biomass Liquefaction: An Overview, in Fundamentals of thermochemical biomass conversion (Eds.: R. P. Overend, T. A. Milne, L.-K. Mudge), Elsevier, London 1985. A. Kruse, A. Gawlik, Ind. Eng. Chem. Res. 2003, 42, 267. J. M. Bouvier, M. Gelus et al., Appl. Energy 1988, 30, 85. Z. Fang, T. Minowa et al., Ind. Eng. Chem. Res. 2004, 43, 2454. Z. Fang, T. Sato, R. L. Smith Jr., H. Inomata, K. Arai, J. A. Kozinski, Bioresour. Technol. 2007, in press. DOI: 10.1016/j.biortech.2007.08.008 G. W. Huber, R. D. Cortright, J. A. Dumesic, Angew. Chem. Int. Ed. 2004, 43, 1549. G. W. Huber, J. N. Chheda, C. J. Barrett, J. A. Dumesic, Science 2005, 308, 1446. B. M. Kabyemela, T. Adschiri, R. M. Malaluan, K. Arai, Ind. Eng. Chem. Res. 1999, 38, 2888. L. Z. Lin, M. Yoshioka et al., Holzforschung 1997, 51, 325. L. Z. Lin, S. Nakagame et al., Holzforschung 2001, 55, 625. L. Z. Lin, Y. G. Yao et al., Carbohydrate Polymers 2004, 57, 123. H. R. Appell, Y. C. Fu, S. Friedman, P. M. Yavorsky, I. Wender, Conversion of Cellulosic Wastes to Oil, Report of Investigation 7560, US Bureau of Mines, Pittsburgh, 1971. [W. J. Connors et al., Holzforschung 1980, 34, 29. A. Vuori, J. B. Bredenberg, Holzforschung 1988, 42, 155. A. Vuori, M. Niemela, Holzforschung 1988, 42, 327. F. P. Petrocelli, M. T. Klein, Ind. Eng. Chem. Prod. Res. Dev. 1985, 24, 635. L. Z. Lin, Y. G. Yao et al., Holzforschung 1997, 51, 316. L. Z. Lin, Y. G. Yao, et al., Holzforschung 2001, 55, 617. A. Demirbas, Energy Convers. Manage. 2000, 41, 633. T. Minowa, T. Kondo et al., Biomass Bioenergy 1998, 14, 517. S. P. Mun, E. B. M. Hassan, J. Ind. Eng. Chem. 2004, 10, 722. A. Krzan, M. Kunaver et al., Acta Chimica Slovenica, 2005, 52, 253.



[33] P. E. Araya, S. E. Droguett et al., Can. J. Chem. Eng. 1986, 64, 775. [34] M. S. El-Gayar, C. A. McAuliffe, Energy Sources 1997, 19, 665. [35] S. P. Mun, E. B. M. Hassan et al., J. Ind. Eng. Chem. 2004, 10, 473. [36] T. Funazukuri, J. S. Cho et al., Fuel 1990, 69, 1328. [37] M. H. Alma, D. Maldas et al., J. of Polym. Eng. 1998, 18, 161. [38] S. H. Lee, T. Ohkita, Wood Sci. Technol. 2003, 37, 29. [39] S. H. Lee, S. Q. Wang, Holzforschung 2005, 59, 628. [40] H. Pan, T. F. Shupe, C.-Y. Hse, J. Appl. Polym. Sci. 2007, 105, 3739. [41] D. Maldas, N. Shiraishi, Biomass Bioenergy 1997, 12, 273. [42] D. G. B. Boocock, K. M. Sherman, Can. J. Chem. Eng. 1985, 63, 627. [43] D. Maldas, N. Shiraishi, Int. J. Polym. Mat. 1996, 33, 61. [44] S. P. Mun, E. B. M. Hassan et al., J. Ind. Eng. Chem. 2001, 7, 430. [45] T. Ogi, S. Yokoyama et al., Chem. Lett. 1985, 8, 1199. [46] M. Yilgin, D. Pehlivan, Energy Convers. Manage. 2004, 45, 2687.


Liquefaction

677

[47] C. Wang, J. Pan, J. Li, Z. Yang, Bioresour. Technol. 2008, 99, 2778. [48] D. Beckman, D. C. Elliott, Can. J. Chem. Eng. 1985, 63 , 99. [49] J. E. Naber, F. Goudriaan, Succesfully using biomass to harness renewable energy in an efficient and cost-effective way, EMIMENT Workshop, Riga 2004. http://www.cpi.umist.ac.uk/ eminent/Confidential/meeting/RigaMeeting/Riga%20Workshop/PresenatieHTUBiofuel.ppt [50] J. E. Naber, F. Goudriaan, HTU®-Diesel from biomass. ACS Division of Fuel Chemistry, 2005. http://membership.acs.org/ P/PETR/2005-Biorefineries/Presentation-08.ppt [51] W. Feng, H. J. van der Kooi, J. de Swaan Arons, Chem. Eng. J. 2004, 98, 105. [52] D. Meier, FVS-Fachtagung „Regenerative Kraftstoffe“, Stuttgart, Germany 2003. [53] E. A. Stadlbauer, Thermokatalytische Niedertemperaturkonvertierung (NTK) von tierische und mikrobieller Biomasse unter Gewinnung von Wertstoffen und Energieträgern im Pilotmaßstab, Final Report of Project 18153, Deutsche Bundesstiftung Umwelt, 2005.
