ABA Process for Microwave Induced Plasma Gasification by Rodolfo Sanchez

CAMPUS MONTERREY

TECHNOLOGICAL STATUS “ABA PROCESS FOR MICROWAVE INDUCED PLASMA GASIFICATION”

MADE BY CENTRO DE CALIDAD AMBIENTAL ITESM CAMPUS MONTERREY

Attention: Ing. Antonio León Sánchez ABA Research S.A. de C.V.

CONFIDENTIAL

Monterrey, N. L.

October, 2011 Ave. Eugenio Garza Sada 2501 Sur Col. Tecnológico C.P. 64849 Monterrey, N.L., México Tel. 8358-2000 y 8358-1400

CAMPUS MONTERREY

Executive Summary Upon request from ABA Research, S.A de C.V., an evaluation was performed on the ABA Process for Microwave Induced Plasma Gasification. For this evaluation, comparative analyses were performed among: foundations of similar technologies, evaluation of the results of experimental tests with different types of materials, computerized modeling, and potential applications for its exploitation. Presently, the technology is in the frontier of state-of-the-art technological developments. The results of the present evaluation are based on experimental tests performed on a medium-scale industrial plant (with capacity to process up to 10 metric tons per day), designed for the realization of pilot tests with raw materials of diverse types. The set of tests performed at such plant during the last 4 years show potentially lower energy consumption when compared to other gasification processes. Considering that the chemistry in the reactions that take place in the process of gasification is the same independently of the process being used, it is to be expected that there will be no significant differences in the composition of the syngas resulting from this process when compared with other technologies. However, the flow (or volume) of syngas would be influenced mainly by the characteristics of the materials to be treated. Several modeling studies following a variety of methods confirm the experimental results that were observed at the present scale, as well as the published results of other research groups. It is recommendable to verify the results of the present process on a continuous process with the purpose of demonstrating experimentally the increase in the real efficiency of the process, and the impact on its economic feasibility. It is to be expected that increases in the efficiency of the process may be achieved as it is scaled to larger capacities. Being this a state-of-the-art technology and realizing its potential, it is highly recommendable to continue the effort of investing in it, in order to reach the status of consolidated technology in the short run.

Ave. Eugenio Garza Sada 2501 Sur Col. Tecnol贸gico C.P. 64849 Monterrey, N.L., M茅xico Tel. 8358-2000 y 8358-1400

CAMPUS MONTERREY

Background ABA Research, S.A. de C.V. is a Mexican company founded in 2006 with the specific purpose of developing new technologies. Its origin dates back in the decade of 1990, when a group of independent Mexican scientists of different specialties, who had previously acted as researchers in different educational institutions undertook research work on this field. By their own initiative, they joined in the search for specific solutions to the production and generation of alternative non-conventional energy based on waste of urban or industrial origin, as well as biomass. Through the years, ABA Research has successfully developed the following products and achievements: 2006-2007 2007 2008 2008-2011 2007 2008 2008 2008 2009

Development of four generations of lab-scale microwave induced plasma gasification reactors. Development of a microwave induced plasma gasifier (Plasmatron). Development, construction and start-up of a microwave induced plasma gasification plant with capacity to process up to 10 tons per day. Development of numerous pilot tests for gasification of diverse raw materials. Patent application presented to IMPI, number MX/A/2007/008317. Patent application presented to PCT, number PCT/MX2008/000081. Patent application presented to US Patent Office, number 20100219062. Patent application presented to European Patent Office, number 08778971.5. Publication of Patent WO2009008693-A1 for the gasification process by WIPO (World Intellectual Property Organization).

The distinctive characteristic of the technology is the use of microwaves as source of energy for the gasification of mostly carbon base materials and with carbon-hydrogen in its molecular structure, which are predominantly converted into a gas stream whose main components are hydrogen and carbon monoxide. This gaseous mixture is known by the name of syngas. This product, the syngas, has a wide range of industrial applications, including: power and steam generation, production of synthetic gasoline and diesel, production of basic petrochemical substitutes and fertilizers, among others.

Foundations • Typical Gasification Process. Gasification is a thermal process that takes place in a reductive atmosphere for the conversion of carbon base materials into a gaseous stream composed predominantly of hydrogen and carbon monoxide, known as syngas. Commonly, the reductive atmosphere is attained by feeding a mass, typically of oxygen, in a stoichiometric relation to the mass of carbon contained in the material to be treated. The global chemical reaction that occurs is a combination of exothermic and endothermic reactions, where the exothermic reaction produces the necessary heat that allows the endothermic reaction to produce the desired mixture of gases according to the following equations: Exothermic Reaction: HC + O2

CO2 + H2O + Heat

(1)

Endothermic Reaction: HC + H2O + Heat

H2 + CO

(2) Ave. Eugenio Garza Sada 2501 Sur Col. Tecnológico C.P. 64849 Monterrey, N.L., México Tel. 8358-2000 y 8358-1400

CAMPUS MONTERREY

Where HC represents the source of carbon • Reformation of crude syngas. In order to improve the ratio of hydrogen and carbon monoxide in the syngas, it is common to add a reactive gas to the process, such as steam, which will react with the incandescent carbon according to the following equation: C + H2O

CO + H2

(3)

• Conventional Gasification In the conventional gasification process, the combustion (exothermic) reactions are controlled by the addition of a stoichiometric amount of air or oxygen such that the combustion supplies only the necessary amount of energy in order to predominantly produce the syngas (endothermic reaction). Typical reaction temperatures for this process fluctuate between 800°C and 1,600°C. The solid residue of the process is ash or slag, and it depends on the temperature of the process being used. Normally, this residue is dispatched to land fields for its final disposal. • Plasma Arc Gasification The Plasma Arc gasification process is a thermal process of high temperature created by an electric arc similar to a torch where a gas is accelerated producing plasma. The typical temperature in a Plasma Arc gasification process fluctuates between 7,000°C and 15,000°C in the zone closest to the torch and is rapidly diffused towards the reactor walls providing a reaction temperature profile which is higher than is required by conventional gasification. Argon is normally used as the plasma gas generator; in certain cases air, oxygen or nitrogen is used for the same purpose, when carrying the process to an industrial scale. Separately to the plasma generator, the process also includes the addition of steam in order to increase the ratio of hydrogen and carbon monoxide in the syngas. The solid residue of the process is a vitrified slag composed of metals encapsulated in a matrix of molten silica. This slag does not produce leach and exceeds the highest standards of environmental control, and can therefore be recycled in the manufacture of different ceramic products including: tile, construction, insulating and decorative materials. • Microwave Induced Plasma Gasification There is little information regarding the phenomenon of microwave induced plasma generation. A research group from the Tokyo Institute of Technology reported in 2003 the process of reformation of hydrocarbons aided by steam plasma generated by a discharge of microwaves (Appendix 1). The labscale experiment used a microwave supply source at a frequency of 2.45 GHz. The experimental basis included measurement equipments in order to identify the generated species responsible for the observed phenomena. For the generation of plasma, the energy from the microwave source generates vibration in the chemical bonds of the steam molecules causing a substantial increase in temperature. There are no reports concerning the temperature reached by the process. Nevertheless, on the basis of modeling studies, it is estimated that it can achieve levels higher than 15,000°C. The chemical species detected in the experiment include free radicals of hydrogen (H.) hydroxyl (OH.) and oxygen (O., O2.). Such species are highly reactive and explain the formation of products similar to those found in other gasification processes, even in the absence of pure oxygen. The research evaluated the reaction of hexane as a model molecule for hydrocarbons, resulting in the production of syngas of high purity. Considering that the temperature profile of microwave induced plasma gasification is similar to or higher than the Plasma Arc or plasma torch, the solid residue of the process would exhibit characteristics similar to those of a Ave. Eugenio Garza Sada 2501 Sur Col. Tecnológico C.P. 64849 Monterrey, N.L., México Tel. 8358-2000 y 8358-1400

CAMPUS MONTERREY vitrified slag. Table 1 shows a comparison of the characteristic features of the conventional, Plasma Arc, and Microwave Induced Plasma Gasification Processes. Table 1. Comparative Table of Different Plasma Generation Processes

Parameter

Conventional Gasification

Arc Plasma

Microwave Plasma

Source Temperature

1,000‐1300 °C

7,000‐15,000 °C

7,000‐15,000°C

Reaction Temperature

1,000‐1,300 °C

Molecular Process

Chemical Oxidation

Physical Acceleration

Vibrational and Physical Acceleration

Plasma Generator

Fuel, O2

Electric Power, air, O2, Ar, N2

Electric Power Water Vapor

Oxidyzing Agent

Air, O2

Water Vapor

Oxidyzing Species

OH., O+ radicals

Moisture Sensitivity

High

Medium

Low

Residue

Slag, Charcol

Vitrified Slag

Energy Consumption Control

Constant

Variable

Description of the ABA Process The continuous reactor for the ABA process is designed to operate at pressures ranging from 1 up to 28 bars, and at temperatures ranging from 1,000°C up to 1,300°C. The distinctive feature of the process lies in the design of the reactor and in the use of microwave plasmatrons. The design consists of two chambers placed one on top of the other. In its standard configuration, the first series of plasmatrons as well as the raw material intakes are located in the higher chamber. As the raw material reacts, it generates a stream composed of carbon monoxide and hydrogen, long hydrocarbon molecules, traces of carbon dioxide, and a solid fraction of incandescent carbon. This stream is transderred to the second chamber equipped with the second series of plasmatrons. In this chamber, super heated steam is added in a stoichiometeric relation. In this section, a reaction of reformation and rectification of the syngas takes place, increasing the ratio of hydrogen and carbon monoxide ensuring the production of gas rectified to syngas. The resulting gas stream from this second chamber is transferred to the cooling section where particles, Ave. Eugenio Garza Sada 2501 Sur Col. Tecnológico C.P. 64849 Monterrey, N.L., México Tel. 8358-2000 y 8358-1400

CAMPUS MONTERREY sulphur, halogen, etc., are removed from the syngas. The solid fraction which does not react to syngas (metals, silicates, etc.) is ejected from the lower part of the reactor as vitrified slag. The sensible heat in these two fractions is recovered as steam.

Applications In the previous sections, a description of the different gasification technologies was reviewed, as well as that corresponding to the ABA Process. In the following paragraphs, several ways of exploitation of the syngas are discussed. • Power Generation. Power generation is probably the first hand application for the exploitation of syngas, considering that existing electric power generation technologies can be adapted to burn syngas or mixtures of this gas with other fossil fuels. Currently, due to the increase in the cost of fossil fuels, manufacturers of equipment designed for electric power generation are focusing their main research efforts to the use of syngas derived from the gasification of low cost fuels such as coal, refinery residues, urban solid waste, and biomass, among others. Syngas has a calorific value that is approximately 50% as that of natural gas. Therefore, twice the volume of syngas would be required in order to generate the same amount of thermal energy to be used in power generation equipments. However, when using a larger volume, the mass of gas increases, resulting in a more efficient transformation to electric energy; resulting in lower costs per unit of electric energy generated and a decrease in greenhouse gas emissions. As an example, General Electric has undertaken research for the design of their turbines in order to use syngas (Appendix 2). For turbines with nominal capacity above 70 MW, their results show that by using syngas, power generation may be increased by as much as 20% as compared to turbines designed for natural gas. Such results show the short term potential for the exploitation of residues which presently are accumulated or confined, and which cause environmental pollution. • Synthetic Fuels. The Fischer-Tropsch (FT) process is a chemical process for the production mainly of liquid hydrocarbons (gasoline, kerosene, gasoil and lubricants) from syngas (CO and H2). It was invented by the German scientists Franz Fischer and Hans Tropsch in the 1920s. During the 1930s, and due to the absence of oil reserves in Germany, the FT process received a strong boost as the source of fuel for warfare machinery. The following major technological leap took place in the 1950s, when oil companies Texaco and Shell began the first plants for the gasification of coal, petroleum coke and mineral oil. Currently, South Africa is the world leader in the production of synthetic fuels based on the huge coal beds the country has. The FT process allows for the production of ultra-low sulphur fuels, as this substance is eliminated in the cleansing process of the syngas. Production of synthetic fuels from gasification of certain kinds of biomass is not economically feasible at low oil prices as those observed in the decades from 1970 to 1990. However, as oil prices soared in the last 10 years investment in these technologies has been considered as especially attractive opportunities. Several studies have been published on the feasibility for producing liquid fuels by FT from biomass. An article published by Tijmensen, et. al. (Appendix 3), shows that for a process of liquid fuels by FT a determined investment is required. Of this, 33% corresponds to the gasification plant, which is made up of Ave. Eugenio Garza Sada 2501 Sur Col. Tecnológico C.P. 64849 Monterrey, N.L., México Tel. 8358-2000 y 8358-1400

CAMPUS MONTERREY the gasifier which represents approximately 18% of the total investment, and the oxygen plant that will supply the oxidizing agent which represents approximately 15% of total investment cost. Technological innovations such as the ABA process for microwave induced plasma gasification can improve the economic feasibility of projects as reported by the quoted source, because they allow improving the energy efficiency of the process, and also because they generally avoid the use of oxygen as an oxidant, which significantly reduces the total investment.

• Environment. Gasification applications for the exploitation of general waste and toxic residues and non-toxic residues which contaminate the environment are endless. The performance of these materials in the gasification process would be the same as the observed by using a carbon based conventional fuel. Materials that can be treated by this process include: hydrocarbon sludge, cleaning sediments from pumping pits, drilling sludge and crude, out of specification crude, spent dispersants, spent oils and lubricants and PCB contaminated oil, among others. In our Working Group, we have developed wide experience in the specific case of exploitation of oil that is contaminated with PCBs by a conventional industrial gasification process (Appendix 4). The results of this research show the production of syngas with a hydrogen content of 45% and 34% carbon monoxide, with an efficiency of removal and destruction of PCBs above 99.99999% and emissions of dioxins and furans below the maximum permissible limit by a factor of 106. Similar results would be expected for the microwave induced plasma gasification process.

Technological Status of the ABA Process Presently, ABA Research has a medium-scale industrial plant (with a capacity of processing up to 10 metric tons per day), designed to undertake tests with different raw materials, and located in the City of Monterrey, NL, Mexico. This plant has been useful for evaluating the conversion potential of diverse materials, including: oil-soaked sawdust, ground tire, sugar cane bagasse, agave bagasse, municipal solid waste, residues from food and paper industries, coal, and petroleum coke, among others. As part of the joint research effort, modeling studies have been performed using tools such as ASPEN Plus, COMSOL, and proprietary software. The results of these studies have confirmed the presence of highly reactive radicals similar to those published by other authors (Appendix 1). Such theoretical results confirm the experimental results found on the formation of syngas and the production of vitrified slag. The phenomenon of microwave induced plasma generation may be considered as a technology in the frontier of state-of-the-art technological developments with a high degree of exploitation of the applied energy. Experimental results have consistently shown lower energy consumption when compared with the calculation of energy requirements obtained by modeling, using the above mentioned tools. Likewise, the experimental results have reported energy requirements noticeable below than those reported by other technologies. The next logical step is the construction of a continuous process plant that confirms the results of the different modeling studies, and of the real process undertaken in the present plant.

Ave. Eugenio Garza Sada 2501 Sur Col. Tecnológico C.P. 64849 Monterrey, N.L., México Tel. 8358-2000 y 8358-1400

CAMPUS MONTERREY

Conclusions and Recommendations This document shows the status of the development of the process of microwave induced plasma gasification and its potential application for the exploitation of carbon based materials. The technology is in the frontier of state-of-the-art technological developments. The results of the present evaluation, based on experimental tests, are indicative of lower energy consumption required to maintain a stable operation of the process in a continuous way. Considering that the chemistry of the reactions that take place in the gasification process is the same, it is expected that no significant differences are to be found in the composition of the syngas resulting from the gasification process. However, the flow (or volume) of syngas would be mainly influenced by the characteristics of the materials to be treated. Modeling studies confirm the observed experimental results, as well as the results of published research groups. It is recommendable to verify the present process results in a continuous process in order to demonstrate experimentally the increase in the real efficiency of the process, and their due impact on economic feasibility. Being, as it is, a technology in the frontier of state-of-the-art technological developments, it is highly advisable to continue the effort to invest in it, in order to reach the status of consolidated technology in the short term.

Prepared by: Porfirio Caballero Mata, Ph.D. Director Environmental Quality Center

Ave. Eugenio Garza Sada 2501 Sur Col. Tecnol贸gico C.P. 64849 Monterrey, N.L., M茅xico Tel. 8358-2000 y 8358-1400

CAMPUS MONTERREY

APPENDIX 1 Hidetoshi Sekiguchi, Yoshihiro Mori. Steam plasma reforming using microwave discharge. Thin Solid Films 435 (2003) 44–48

Ave. Eugenio Garza Sada 2501 Sur Col. Tecnológico C.P. 64849 Monterrey, N.L., México Tel. 8358-2000 y 8358-1400

Thin Solid Films 435 (2003) 44–48

Steam plasma reforming using microwave discharge Hidetoshi Sekiguchi*, Yoshihiro Mori Tokyo Institute of Technology, Department of Chemical Engineering, 2-12-1 O-okayama, Meguro-ku, Tokyo 152-8552, Japan

Abstract The purpose of the study was to prepare atmospheric pure-steam plasma using a microwave discharge and to apply the plasma to the reforming of hydrocarbon for hydrogen production. The experiment was conducted with 2.45-GHz microwave power supply with a special wave-guide designed elsewhere. The results showed that the pure steam plasma could be stably sustained, and analysis of emission spectra indicated that O, H and OH existed in the plasma. Hexane, as a model of gasoline, was used for the hydrocarbon reforming experiment. The results showed that H2 and CO were predominantly produced, as suggested from equilibrium calculations, and that reforming could be accomplished. The conversion of hexane and steam was affected by the plasma power, and the ratio and total flow rate of the reactants. A fuel-cell power system including the steam plasma reformer was evaluated from experimental data and the results suggested that improvements are required for practical use of the plasma reforming method. 䊚 2003 Elsevier Science B.V. All rights reserved. Keywords: Steam plasma; Microwave discharge; Reforming; Hydrogen

1. Introduction Recently, fuel cell systems have attracted much attention for highly efficient and decentralized electric power sources. This power source can be widely applied in commercial and military uses. In particular, the installation of such systems in vehicles is being competitively developed in many major motor companies. A major concern in developing fuel-cell systems is how to generate hydrogen as fuel. The reformation of various hydrocarbons, such as methanol, methane and gasoline, are being considered as hydrogen sources. Every hydrocarbon source has advantages and disadvantages from the point of view of reforming temperature, reaction rate, storage system, supply infrastructure, etc. Conventional reforming is carried out thermally with steam and oxygen. Partial oxidation of a hydrocarbon takes place to provide reaction heat, because the reforming reaction with steam is endothermic. Noble metal catalysts are usually required to enhance the reaction rate. Therefore, the reforming system is easily affected by the deactivation of catalysts caused by impurities in the hydrocarbon and carbon deposits. One attractive method for reforming hydrocarbons is to use plasma w1,2x. The plasma contains highly active *Corresponding author.

species, such as electrons, ions and radicals, which may enhance the reforming reaction rate. High reactivity also eliminates the need for catalysts in the system. These advantages, as well as its high energy density, lead to compactness of the plasma reformer. Moreover, the plasma system can be adapted for various hydrocarbons, including natural gas, gasoline, heavy oils and biofuels. A fast response time can be also achieved because the plasma is operated by electricity. However, the utilization of electricity seems a disadvantage from the viewpoint of energy efficiency. Microwave discharge is one technique used to obtain a non-equilibrium plasma, even under atmospheric pressure, at which the electron temperature is approximately 4000–6000 K, while the heavy particle temperature is around 2000 K w3x. Discharge techniques under atmospheric pressure have been extensively studied elsewhere w4–8x. The purification of noble gases w9x and the decontamination of radioactive wastes w10x have been studied using atmospheric microwave plasma. When steam is used as the plasma supporting gas in microwave discharge, radicals such as H, OH and O are generated, as well as high-energy electrons, and both reductive and oxidative conditions are provided in the plasma, indicating that steam plasma is effective for various treatments of materials.

H. Sekiguchi, Y. Mori / Thin Solid Films 435 (2003) 44â&#x20AC;&#x201C;48

exists. The equilibrium calculation suggests that reforming should be carried out above a ratio of unity, which is the stoichiometric condition indicated by Eq. (1). 3. Experimental

Fig. 1. Equilibrium composition for C6H14 and H2O. Initial amounts of C6H14 and H2O were 1 and 6 mol, respectively (OyCs1).

The purpose of this study was to prepare atmospheric pure steam plasma using microwave discharge and to apply the plasma to the reforming of hydrocarbon for hydrogen production. The reforming of n-hexane as a model of gasoline was experimentally carried out using steam plasma. System analysis of a fuel-cell power source including the plasma reformer was also carried out using the experimental data. 2. Equilibrium compositions

The experimental apparatus is shown in Fig. 3. The experiment was conducted with a 2.45-GHz microwave power supply having maximum power of 2.8 kW. The wave-guide used was as the same as that designed for other work w5x and which can efficiently generate plasma columns having high electron density. As mentioned above, n-hexane was used as a model of gasoline for hydrocarbon reforming. Steam and hexane were heated and introduced into a quartz tube reactor having a diameter of 12 mm and a length of 500 mm. The reactor was inserted in the wave-guide, where electromagnetic field was concentrated. Tangential injection of the feed enabled stabilization of the steam plasma. The composition of the product gas was analyzed by gas chromatography (GC). Analysis of the emission spectra was also carried out. The experimental conditions are summarized in Table 1. 4. Results and discussion Emission spectroscopy indicated that spectra originated from OH, O and H were observed in the pure steam plasma, indicating the dissociation of steam. The emissions of C and C2 were detected when hexane was injected. The effect of power applied on product compositions is shown in Fig. 4, for which the feed rates of C6H14

Equilibrium compositions have been calculated for a preliminary investigation into the reforming, even though the microwave plasma is classified as nonthermal plasma and high-energy electrons are involved in the reactions in the plasma. Calculations were performed for a mixture of steam and n-hexane by minimization of the Gibbs free energy, considering typical species that contain C, H or O elements w11,12x. The results are shown in Fig. 1, in which the CyO molar ratio in the feed gas was equal to unity. In the temperature range between 1000 and 2000 8C, H2 and CO become dominant, while H2 dissociates into atoms with increasing temperature. The stoichiometric reaction can be written with the reaction enthalpy as follows: C6H14q6H2Oâ&#x201E;˘6COq13H2 DHs1249 kJymol

(1)

Fig. 2 shows the effect of the OyC ratio on the composition at 1500 8C. Solid carbon exists up to Oy Cs1, at which both H2 and CO increase with the ratio. A slight decrease in CO and increases in both H2 and CO2 are observed above OyCs1, at which excess H2O

Fig. 2. Effect of OyC ratio on equilibrium composition. Initial amount of C6H14 was 1 mol; Ts1500 8C.

H. Sekiguchi, Y. Mori / Thin Solid Films 435 (2003) 44–48

Fig. 3. Experimental apparatus.

and H2O were fixed. The product gas contains predominantly H2, CO and H2O, and the reforming process proceeds in the steam plasma. The amounts of H2 and CO increase with power, while H2O decreases. Slight formation of solid carbon was estimated from GC analysis. A small amount of carbon dioxide was also detected in the product gas. Fig. 5 shows the conversions of C6H14 and H2O defined by: Conversion ŽC6H14.s

Conversion ŽH2O.s

FC6H14yRC6H14 FC6H14

and H2O show the same behavior as H2 production, as shown already. However, complete reforming of hexane was not attained. The effect of steam feed rate on conversion is shown in Fig. 6, for which the C6H14 feed rate was fixed.

(2)

FH2OyRH2O FH2O

(3)

where F and R denote the flow rates of the feed and the product gas, respectively. The conversions of C6H14 Table 1 Experimental conditions Plasma torch Inner diameter Length Feed flow rate C6H14 H 2O Molar ratio in the feed (OyC) Power

12 mm 500 mm 0.37–1.00 mmolys 4.86–12.00 mmolys 1–3 1.6–2.5 kW

Fig. 4. Effect of power applied on product composition. Feed rates of C6H14 and H2O were 0.81 and 9.72 mmolys, respectively (OyCs2).

H. Sekiguchi, Y. Mori / Thin Solid Films 435 (2003) 44â&#x20AC;&#x201C;48

Fig. 5. Effect of power applied on conversion. Feed rates of C6H14 and H2O were 0.81 and 9.72 mmolys, respectively (OyCs2).

Hence, the increase in ratio implies an increase in steam feed rate. The conversion of C6H14 shows a peak for a ratio of approximately 1.5, while the conversion of H2O decreases with increasing ratio, i.e. the steam feed rate. One reason for the maximum may be ascribed to the temperature distribution in the reactor, as well as insufficient mixing of steam and hexane. Further investigation will explain these results. When the total feed rate was changed with both the power and the OyC ratio set constant, the conversions of C6H14 and H2 decreased with total flow rate. Moreover, a high C6H14 feed rate enhanced the conversion of H2O, but reduced the conversion of C6H14 itself.

WFyWP WF

(4)

where WP and WF are the power consumed by the plasma reformer and generated by the fuel cell stack, respectively. The numerator indicates energy available from the system. The energy required in the heaters (QH2O and QC6H14) and the shift reactor is not considered, because heat can be recovered from the exhaust gas of the shift reactor. The generation of H2 by the shift reactor is also added, assuming that the shift reactor completely converts CO into CO2 without additional heat. The calculation was performed with experimental data for which the highest efficiency for H2 production (H2 production rate divided by plasma power) was achieved. The results are indicated in Table 2, with another case for which complete reforming of hexane was assumed. The performance based on the experimental data is too low to supply electric power because the magnetron efficiency is excluded in the estimation: a typical efficiency of up to 80% leads to negative performance of the system. Therefore, improvements are required to achieve complete conversion of hexane. Further evaluation with real data for each element in the system will elucidate the effectiveness and requirements for the steam plasma reformer. 6. Conclusion Pure steam plasma could be stably obtained using microwave discharge under atmospheric pressure. The plasma contained radicals such as OH, O and H, which were detected by emission spectroscopy. The plasma was applied to hexane reforming and the results showed that the production of hydrogen could be achieved;

5. Evaluation of fuel cell system The plasma reformer proposed here consumes electric power, and hence the performance of a fuel-cell power system including steam plasma reforming was evaluated from the viewpoint of energy efficiency. The system proposed is shown in Fig. 7. Two heaters were used for the evaporation of both H2O and C6H14. The plasma reformer investigated in this research provides hydrogen for fuel cells. The shift reactor converts CO into CO2 with H2O and generates H2 simultaneously. The separator should be equipped due to the incomplete reaction of C6H14 observed in the experiment. However, the separator can be eliminated if C6H14 is reacted to completion. In the evaluation, the separator is not taken into account. The fuel cell stack generates electric power. The type of the fuel cell is not specified and ideal efficiency for the cell is assumed to evaluate system performance. The performance is evaluated according to:

Fig. 6. Effect of steam flow rate on conversion. Feed rate of C6H14 was 0.81 mmolys; power was 1.8 kW.

H. Sekiguchi, Y. Mori / Thin Solid Films 435 (2003) 44â&#x20AC;&#x201C;48

Fig. 7. Fuel cell power system with plasma reformer. Table 2 System performance evaluated

Data from experiment Complete reforming

References

Conversion of C6H14

WP (W)

WF (W)

0.66 1.0

1800 1800

2173 3647

0.17 0.51

Feed rates of C6H14 and H2O were 0.81 and 6.08 mmolys, respectively.

however, the conversion of hexane was not sufficient. A fuel-cell power system including the plasma reformer was evaluated and the performance was insufficient using the experimental data, showing that improvements are required for the plasma reforming technique. Acknowledgments This research was supported by the Ministry of Education, Science, Sports and Culture under Grant-inAid for Scientific Research (B) No 13558054.

w1x L. Bromberg, D.R. Cohn, A. Rabinovich, Int. J. Hydrogen Energy 22 (1997) 83. w2x L. Bromberg, D.R. Cohn, A. Rabinovich, N. Alexeev, A. Samokhin, R. Ramprasad, S. Tamhankar, Int. J. Hydrogen Energy 25 (2000) 1157. w3x M.D. Calzada, M. Moisan, A. Gamero, A. Sola, J. Appl. Phys. 80 (1996) 1. w4x J. Hubert, M. Moisan, A. Ricard, Spectrochim. Acta B 33 (1979) 1. w5x M. Moisan, Z. Zakrrzewski, R. Pantel, P. Leprince, IEEE Trans. Plasma Sci. 12 (1984) 203. w6x M. Moisan, Z. Zakrrzewski, J. Phys. D 24 (1991) 1025. w7x M. Moisan, G. Sauve, Z. Zakrrzewski, J. Hubert, Plasma Sources Sci. Technol. 3 (1994) 584. w8x M. Moisan, Z. Zakrrzewski, R. Etemadi, J.C. Rostaing, J. Appl. Phys. 83 (1998) 5691. w9x J.C. Rostaing, F. Bryselbout, M. Moisan, J.C. Parent, C.R. Acad. Sci. Paris 1-IV (2000) 99. w10x H.F. Windarto, T. Matsumoto, H. Akatsuka, K. Sakagishi, M. Suzuki, J. Nucl. Sci. Technol. 37 (2000) 787. w11x H. Sekiguchi, T. Honda, A. Kanzawa, Proc. ISPC 10 (1991) 1.5-2. w12x G. Eriksson, Acta Chem. Semicond. 25 (1971) 2651.

CAMPUS MONTERREY

APPENDIX 2 General Electric Company. Syngas Turbine Technology. GEA18028 (09/2010)

Ave. Eugenio Garza Sada 2501 Sur Col. Tecnol贸gico C.P. 64849 Monterrey, N.L., M茅xico Tel. 8358-2000 y 8358-1400

GE Energy

Syngas Turbine Technology

Turbine Innovation Realized 2x7F Syngas Coal IGCC (USA)

GE turbines for syngas and low-Btu fuel

2010 2x9E Steel Mill (Asia)

applications are operating at locations around the world with more than one million hours of total operation. This vast experience covers operations employing GE and other gasification technologies, and a variety of fuels including high- and low-sulfur coals and

2x6F Syngas Refinery (Asia)

3x9E Refinery (Europe)

technology for syngas applications.

3x9E Steel Mill (Europe)

1x7FA IGCC (USA)

1990

GE Turbines On Low-Btu Fuels

edge in providing customers with wellproven and experienced combined-cycle

2000

2x6FA Refinery (USA)

1x7FA IGCC (USA)

petroleum coke. GE offers an unparalleled

S Y N G A S T U R B I N E T E C H N O LO GY

Fuel Flexible Solutions Sources for generating power are becoming more varied, and more stringent emission requirements are fueling a need for flexible solutions to meet growing energy demand. At GE, we’re developing solutions today that are flexible enough to integrate into your diverse portfolio of power generating options, helping you to profitably guide the industry—and your community—into the future. As a leader in combined-cycle gas turbine technology, GE has invested its time, resources and expertise to develop a range of efficient, reliable gas turbines to help energy providers meet these new challenges. GE’s versatile gas turbines can operate on a variety of low-Btu fuels, in a wide variety of power applications, including hydrogen, low-rank steel industry furnace gases, light distillates, heavy residuals from refining and syngas. Our solutions can help customers enhance fuel utilization, reducing fuel costs and improving revenues.

Syngas Gas Power Generation in a Carbon Constrained World Many utilities look towards abundant supplies of relatively low-cost coal for power generation. Increasingly the use of coal—and the emissions and carbon burning it produces—is coming under scrutiny, and many power generators are turning to integrated gasification combined-cycle (IGCC) technology. At the front end of IGCC is gasification, which takes a carbon-based fuel source such as coal, refinery residuals or biomass, and under high heat and pressure, converts it into a synthesis gas (or syngas) comprised of H2 and CO. Impurities and carbon can be removed easily and economically from the syngas stream on a pre-combustion basis—leaving a hydrogen-rich fuel—before it is burned to create electricity. Carbon capture technologies, which have been in commercial operation for many years, offer the ability to efficiently and cost effectively remove carbon from syngas for permanent storage or use in enhanced oil recovery before burning the fuel. The resulting gas is essentially a carbon-free, high hydrogen fuel available for combustion within a combined-cycle power plant. GE’s syngas power turbines offer a powerful, reliable solution that can operate on the high hydrogen fuels resulting from this carbon separation. While today’s GE syngas turbines have been used successfully on fuels with up to 50% hydrogen, we continue to advance the capabilities of the next generation of gas turbines for high hydrogen fuels. We’re working on breakthroughs that will deliver cleaner, efficient technologies to customers who may capture carbon for storage.

SYNGAS TURBINE TECHNOLOGY

Challenges Met, Innovation Attained Our syngas turbine platforms are built upon GEâ&#x20AC;&#x2122;s extensive experience and rigorous engineering processes. Advances in technology will deliver high efficiency and reliability. Among these advances is the Multi Nozzle Quiet Combustion (MNQC). One of the key challenges of hydrogen fuel is the high flame speed. The MNQC, a diffusion flame-based system that is free of the sensitivities to flame speed and combustion instability (combustion dynamics) that are inherent to lean pre-mix combustion systems was developed to address this challenge. GEâ&#x20AC;&#x2122;s MNQC system has been designed to run on low-Btu fuels and is capable of operating on many varieties of syngas, including high H2 fuels. It is built for high efficiency and offers superior benefits for customers operating in base load applications. This combustor is also capable of operating on natural gas for start-up and shut-down operations, and can operate at base load on natural gas for extended periods of time if syngas is not available. With MNQC technology, GE can offer a similar combustor configuration for several turbine products using syngas and high hydrogen fuels, all based upon a combustion system design that has been in use since the 1990s, and can operate with un-shifted and shifted (hydrogen-rich) syngas.

Syngas Turbine Fuel Applications

7EA 9E 6FA 7F 9F FUEL HEATING VALUE LOW Air IGCC Syngas Blast Furnace Gas

S Y N G A S T U R B I N E T E C H N O LO GY

MEDIUM O2 IGCC Syngas GTL Off-gas

HIGH High H2 for CCS High H2 for EOR

SYNGAS TURBINE TECHNOLOGY

Gas Turbines for IGCC Syngas Applications1 Gas Turbines

Combined-Cycle

Model

Nominal Syngas Power Rating2

Model

Nominal Syngas Output Power3

46 MW (50/60 Hz)

106B

70 MW (50/60 Hz)

7EA

80 MW (60 Hz)

107EA

120 MW (60 Hz)

140 MW (50 Hz)

109E

210 MW (50 Hz)

6FA

92 MW (50/60 Hz)

106FA

140 MW (50/60 Hz)

7F Syngas

232 MW (60 Hz)

207F Syngas

710 MW (60 Hz)

9F Syngas

286 MW (50 Hz)

209F Syngas

880 MW (50 Hz)

Notes: (1) Conventional gasification fuel, without CO2 capture. (2) Performance at ISO conditions. (3) No integration with process. Steam turbine and generator product fit TBD. Assumes multishaft configuration.

Gas Turbines for Syngas Applications 85 MW

SIMPLE CYCLE OUTPUT (MW)

85 MW

7EA

NATURAL GAS 77 MW

SYNGAS

92 MW

6FA 126 MW

140 MW

9E 187 MW

232 MW

7F 256 MW

S Y N G A S T U R B I N E T E C H N O LO GY

285 MW

GE Syngas Turbine Solutions GE’s portfolio of syngas capable turbines, includes units for both 50 Hz and 60 Hz segments, simple-cycle configurations with output ranging from 46–300 MW, and combined-cycle configurations with output ranging from 70–880 MW, depending on fuel and site specific conditions. 6B…Reliable and rugged 50/60 Hz power This rugged and reliable gas turbine, a popular choice for mid-range power generation service, has a well-documented availability of 94.6% and 99% reliability. With over 1,100 units worldwide, the dependable 6B features low capital investment and low maintenance costs. It has accumulated over 60 million operating hours in a wide range of applications—including simple-cycle, heat recovery, combined-cycle, and mechanical drive. Introduced in 1978, many upgrades are available to improve the performance of earlier versions, including rotor life extension and combustion system retrofits that can deliver 5 ppm NOx when operating on natural gas. With its lengthy industrial experience and high reliability, the 6B is an excellent fit for industrial and oil and gas applications, providing horsepower and high exhaust energy. The 6B has long operating experience on a variety of medium- or low-BTU fuels, including syngas produced from oil and steel mill gasses.

7EA…Proven performance for 60 Hz applications The size of the versatile 7EA gas turbine enables flexibility in plant layout and fast, low-cost additions of incremental power. With high reliability and availability, this unit provides strong efficiency performance in simple-cycle and combined-cycle applications—and is ideally suited for power generation, industrial, mechanical drive, and cogeneration applications. 7E/EA units have accumulated millions of hours of operation using crude and residual oils, and were featured in the first large scale IGCC demonstration plant at Coolwater that operated in the mid 1980s.

9E…Flexible and adaptable performance for 50 Hz applications Since its introduction in 1978, GE’s 9E gas turbine fleet of 430+ units has accumulated over 22 million hours of utility and industrial service—often in arduous climates ranging from desert heat and tropical humidity to arctic cold. Capable of operating on a variety of medium- or low-BTU fuels, including syngas produced from oil and steel mill gasses, the 9E is a quick power solution also well suited for IGCC or mechanical drive applications. This reliable, low first-cost machine has a compact design that provides flexibility in plant layout and easy addition of incremental power when phased capacity expansion is required.

6FA (50/60 Hz)…Advanced technology mid-sized combined-cycle and cogeneration With over 2.5+ million operating hours and more than 110 units installed or on order, the 6FA gas turbine is a good fit for local power for industrial complexes. It delivers high efficiency and high availability, and provides the operating flexibility needed for harsh environments. A direct down-scaling of the proven 7FA, the 6FA is rated at 85–95 MW when operating with syngas. Four 6FAs have been operating syngas at two refinery cogeneration plants since the early 2000s.

7F Syngas…Large baseload syngas performance for 60 Hz Building on GE’s F-class syngas experience, the 7F Syngas turbine was developed specifically for syngas operation. The configuration of the 7F Syngas turbine combines materials with known syngas compatibility and turbine components designed for the increased mass flow in syngas applications. These elements allow the 7F Syngas turbine to generate 232 MW on dry syngas at ISO conditions. This turbine—like GE’s other syngas turbines—is capable of operating on syngas and on high hydrogen (carbon captured) fuels. The first two units shipped to Duke Energy’s Edwardsport IGCC plant in 2010.

9F Syngas…Advanced turbine technology for 50 Hz applications The latest addition to GE’s syngas turbine portfolio is the 9F Syngas turbine. This unit, based on the 9FA, building on the world’s most experienced fleet of highly efficient 50 Hz large units. The 9F Syngas turbine can be arranged in a single-shaft or multi-shaft configuration that combines one or two gas turbines with a single steam turbine. This turbine is capable of operating on syngas (non-carbon captured) or a high hydrogen fuel (carbon capture).

SYNGAS TURBINE TECHNOLOGY

Refinery Residuals for Cogeneration Gasification technology combined with GE’s syngas turbines is an effective way to use refinery residuals and generate N2, H2, steam and power in a petrochemical complex. GE deploys its advanced gas turbine technology to deliver greater performance levels than ever before, offering customers gas turbine solutions with a wide range of fuel and process integration flexibility. Based on its significant gas turbine and syngas experience, GE is pleased to offer solutions that deliver high efficiency and reliability for advanced IGCC and cogeneration plants.

Plant Integration Developing an IGCC plant or gasifier in a petrochemical complex is a capital-intensive project, so delivering results is important. Seamlessly integrating the gasification and power islands will help enable operators of IGCC and cogeneration plants to derive return on their investment. Our syngas turbines provide air extraction (air from the gas turbine compressor) to the process plant, allowing for a reduction in the number of compressors required to supply air to gasifier, air separation unit, or other plant-level process, delivering value to the operators.

GE offers expertise in the integration of the syngas turbine, including:

High P, T Air to reduce compressor loads

• Steam-side integration • Nitrogen return • Full steam and air integration, including air extraction and nitrogen return

Syngas or high H2 N2 from Air Separation Unit for dilution Process heat for fuel heating

S Y N G A S T U R B I N E T E C H N O LO GY

SYNGAS TURBINE TECHNOLOGY

Flexibility of GE Syngas Turbine Technology GE leads the world in the application of its heavy duty gas turbines to gasification combined-cycle gas projects. Our success with low- and medium-Btu fuel gases is a consequence of extensive full-scale laboratory testing on various fuels for nearly 24 years at GE’s combustion laboratory in Schenectady, New York, and, since 2002, testing at the combustion development laboratory in Greenville, South Carolina.

In these facilities, we develop and validate system components and the impacts of impact of various characteristics on performance of a combustion system. At our Greenville test facility, for example, we are able to validate the 7F syngas turbine combustion system, testing it at full pressure, temperature and flows. The facility also has the capability to blend a variety of syngas-like fuels. We can also test the turbine on start-up fuel (natural gas) at full speed, no load conditions. Typical test campaigns examine combustion full and part load performance, including combustion dynamics, emissions and exit profile. In addition, thermocouples, strain gauges and thermal paint – combined with advanced computational fluid dynamics and finite element analysis – allow full durability validation of the system. Combustion lab testing has evaluated performance over a range of fuel space and load points. Results from these tests have confirmed that the system will be able to meet customers’ performance goals.

GE Delivers GE Energy provides innovative, technology-based products and service solutions across the full spectrum of the energy industry and is committed to investing in a cleaner, smarter, and more efficient future. To put GE’s proven syngas turbine technology to work for you, contact your local GE representative or visit http://www.ge-energy.com

S Y N G A S T U R B I N E T E C H N O LO GY

SYNGAS TURBINE TECHNOLOGY

CAMPUS MONTERREY

APPENDIX 3 Michiel J.A. Tijmensen, Andre P.C. Faaij, Carlo N. Hamelinck, Martijn R.M. van Hardeveld. Exploration of the possibilities for production of Fischer Tropsch liquids and power via biomass gasification. Biomass and Bioenergy 23 (2002) 129 – 152

Ave. Eugenio Garza Sada 2501 Sur Col. Tecnológico C.P. 64849 Monterrey, N.L., México Tel. 8358-2000 y 8358-1400

Biomass and Bioenergy 23 (2002) 129 – 152

Exploration of the possibilities for production of Fischer Tropsch liquids and power via biomass gasi%cation Michiel J.A. Tijmensena , Andr,e P.C. Faaija; ∗ , Carlo N. Hamelincka , Martijn R.M. van Hardeveldb a Department b Shell

of Science, Technology and Society, Utrecht University, Padualaan 14, 3584 CH, Utrecht, Netherlands Global Solutions International BV, Badhuisweg 3, P.O. Box 38000, 1030 BN, Amsterdam, Netherlands

Received 17 September 2001; received in revised form 22 February 2002; accepted 13 March 2002

Abstract This paper reviews the technical feasibility and economics of biomass integrated gasi%cation–Fischer Tropsch (BIG-FT) processes in general, identi%es most promising system con%gurations and identi%es key R&D issues essential for the commercialisation of BIG-FT technology. The FT synthesis produces hydrocarbons of di;erent length from a gas mixture of H2 and CO. The large hydrocarbons can be hydrocracked to form mainly diesel of excellent quality. The fraction of short hydrocarbons is used in a combined cycle with the remainder of the syngas. Overall LHV energy e?ciencies,1 calculated with the @owsheet modelling tool Aspenplus , are 33– 40% for atmospheric gasi%cation systems and 42–50% for pressurised gasi%cation systems. Investment costs of such systems (367 MWth ) are MUS$ 280 – 450,2 depending on the system con%guration. In the short term, production costs of FT-liquids will be about US$ 16=GJ. In the longer term, with large-scale production, higher CO conversion and higher C5+ selectivity in the FT process, production costs of FT-liquids could drop to US$ 9=GJ. These perspectives for this route and use of biomass-derived FT-fuels in the transport sector are promising. Research and development should be aimed at the development of large-scale (pressurised) biomass gasi%cation-based systems and special attention must be given to the gas cleaning section. ? 2002 Elsevier Science Ltd. All rights reserved. Keywords: Biomass; Gasi%cation; Fischer Tropsch synthesis; FT-liquids; Polygeneration; Diesel; Combined cycle

1. Introduction 1.1. General background To prevent climate change induced by human activity, greenhouse gas emissions must be dramatically ∗ Corresponding author. Tel.: +31-20-353-7643; fax: +31-30253-7601. E-mail address: a.faaij@chem.uu.nl (A.P.C. Faaij). 1 E?ciency throughout this paper is on LHV wet basis, unless indicated otherwise. 2 All Cost numbers are in US$ 2000 .

reduced. Renewable energy (e.g. solar, wind and biomass) could play a major role in achieving this. Biomass is a renewable energy source when carbon dioxide emissions caused by its use are absorbed by newly grown biomass. Only biomass o;ers the possibility to produce liquid, carbon neutral, transportation fuels. Ethanol, methanol and synthetic hydrocarbons, as well as hydrogen can be produced from biomass and could o;er feasible alternatives for the transport sector on foreseeable term [1,2]. This is particularly relevant since transport is responsible for a large part of global CO2 emissions. The global trend is that

130

M.J.A. Tijmensen et al. / Biomass and Bioenergy 23 (2002) 129 – 152 Recycle loop

for unconverted syngas (optional)

Pre-treatment:

Gasification:

Gas Cleaning:

Gas Processing

FT synthesis:

- air or oxygen - pressurised or atmospheric - direct/indirect

‘wet’ cold or ‘dry’ hot

grinding drying

feedstock is poplar wood

reforming, optional CH4 → H2+CO shift, optional (adjusting the H2/CO ratio) CO2 removal (reducing amount of inert) optional

Gas turbine

reactor type Off gas Slurry or fixed bed Power FT liquids

Fig. 1. A basic schematic view of the key components for converting biomass to FT-liquids combined with gas turbine (combined cycle) power generation.

the share of transport in the total energy consumption is increasing, especially in developing countries [3]. Some recent studies indicated that the use of Fischer-Tropsch (FT) technology for biomass conversion to synthetic hydrocarbons may o;er a promising and carbon neutral alternative to conventional diesel, kerosene and gasoline [2,4]. The FT process is a process capable of producing liquid hydrocarbon fuels from syngas. The recent interest in FT synthesis has grown as a consequence of environmental demands, increased use of natural gas from remote locations and technological developments. First, FT-liquids are totally free of sulphur and contain very few aromatics compared to gasoline and diesel, which results in lower emission levels when applied in internal combustion engines. Known reserves of natural gas have increased but a signi%cant portion has been assigned ‘stranded’ [5]. Conversion on location of natural gas into shippable hydrocarbon liquids is possible by syngas generation followed by FT synthesis. This is demonstrated on full commercial scale by Shell in Malaysia. Products made by the FT synthesis, hydrocarbons of di;erent length, can be transported by the same means as oil. Shell (natural gas based) and Sasol (coal based) apply FT synthesis on commercial scale. This growing market for FT technology also drives further technological development of this type of process. FT synthesis from biomass-derived syngas, however, has received little attention so far.

1.2. Rationale In principle, numerous process con%gurations for the conversion of biomass to FT-liquids are possible, e.g. depending on the gasi%er types, the FT process and the gas cleaning process considered. A scheme of the main process steps to convert biomass to FT-liquids (and power) and possible variations is shown in Fig. 1. Di;erent gasi%cation methods, covering atmospheric and pressurised, air-blown and oxygen-blown, indirect and direct, can produce a wide range of syngas compositions, with H2 =CO ratios varying between 0.45 and 2. Any raw-biomass-derived syngas contains contaminants like H2 S, NH3 , dust and alkalis. Consequently, the syngas needs to be cleaned and processed to make it suitable for the FT synthesis. Several processing steps (like reforming and shift) can be applied to manipulate the syngas composition prior to the FT reactor. The FT synthesis can be realised in di;erent reactor types. Furthermore, o;gas from the FT synthesis can either be recycled partially (full conversion mode) or used directly in a gas turbine for electricity production (once through mode). Last but not least, economies of scale are important for this type of technology [e.g. 2]. Some process components may be more suited for upscaling than others, which may lead to di;erent ‘optimal’ technology for di;erent capacities. In total, all those variables lead to a large number of possible process con%gurations to produce FT-liquids from biomass.

M.J.A. Tijmensen et al. / Biomass and Bioenergy 23 (2002) 129 – 152

1.3. Objectives The main objective of this study is to evaluate the di;erent options to use biomass for the production of FT-liquids. The main research questions are: • To explore the technical feasibility and economics of biomass integrated gasi%cation—Fischer Tropsch (BIG-FT) processes in general, with speci%c attention for gas cleaning requirements. • To identify most promising system con%gurations; various biomass gasi%cation processes will be studied in combination with FT-concepts in two main categories: 1. Full conversion FT with the possible use of a gas turbine, focussed on a maximum amount of FT-liquids. 2. Once through FT with co-%ring the o;gas in a gas turbine. • To investigate economies of scale of BIG-FT conversion concepts. • To explore the technical and economic perspectives of this route on the longer term. • To identify key R&D issues for the commercialisation of BIG-FT technology. 1.4. Methodology The work consists of several steps: %rst, a technology assessment on gasi%cation, gas cleaning, syngas processing, FT conversion and combined cycle is made. Besides information from literature, experts from technology manufactures and research institutes were consulted to identify the potential problems with the use of FT processes for biomass-derived syngas. Manufacturers have been consulted for process data. The assessment includes some technologies that are not applied commercially at present, such as advanced high temperature gas cleaning options. Second, promising system con%gurations were selected for further performance modelling with help of the @owsheeting program Aspenplus . Aspenplus is used to calculate energy and mass balances. Third, an economic evaluation is performed. Again, manufacturers have been consulted for cost data of various components. Fourth, an extensive sensitivity analysis is performed, including economies of scale of BIG-FT systems. And %nally, the various system con-

131

%gurations are compared, conclusions are drawn and recommendations on R&D issues are formulated. 2. System description 2.1. The FT process 2.1.1. Reaction mechanism and selectivity The FT reaction produces hydrocarbons of variable chain length from a gas mixture of carbon monoxide and hydrogen. Nowadays, this process is operated commercially at Sasol South Africa (from coal-derived syngas) and Shell Malaysia (from natural gas-derived syngas). The main mechanism of the FT reaction is CO + 2H2 → –CH2 – + H2 O; ◦ RHFT = −165 KJmol−1 :

(1)

The –CH2 – is a building stone for longer hydrocarbons. A main characteristic regarding the performance of the FT synthesis is the liquid selectivity of the process. The liquid selectivity is determined by the so-called ‘chain growth probability’. This is the chance that a hydrocarbon chain grows with another CH2− group, instead of terminating. The products made by the FT reaction are hydrocarbons of di;erent length. A high liquid selectivity (or C5+ selectivity: SC5+ ) is necessary to obtain a maximum amount of long hydrocarbon chains. The yield in the C1 –C4 range decreases with increasing SC5+ ; any C1 –C4 in the o;gas may e?ciently be used in a gas turbine for power generation. The relation between the hydrocarbon yield and the chain growth probability is described by the Anderson–Schulz–Flory (ASF) distribution [4,5,7]. The ASF distribution describes the molar yield in carbon number as: fraction Cn = n−1 (1 − ), where is chain growth probability and n the length of the hydrocarbon, which makes 1 − the chance that the chain growth terminates. Fig. 2 shows the hydrocarbon yields for di;erent values of chain growth probability. Selectivity is in@uenced by a number of factors, either catalyst dependent (type of metal (iron or cobalt), support, preparation, pre-conditioning and age of catalyst) or non-catalyst dependent (H2 =CO ratio in the feed gas, temperature, pressure and reactor type).

132

M.J.A. Tijmensen et al. / Biomass and Bioenergy 23 (2002) 129 – 152

Fig. 2. Product distribution for di;erent for FT synthesis [31].

Fig. 4. Slurry bed FT reactor [12].

Fig. 3. Tubular %xed bed FT reactor [12].

The division in iron and cobalt is relevant because the water–gas-shift reaction, takes place only signi%cantly over an iron catalyst. The FT synthesis uses H2 and CO at a ratio near 2:1:1, depending on selectivity. Since biomass gasi-

%cation in most cases leads to a signi%cantly lower H2 =CO ratio in the feed gas, a shift reaction may be necessary. The FT process is generally operated at pressures ◦ ranging from 20 to 40 bar and at 180 –250 C. Higher partial pressures of H2 and CO lead to higher liquid selectivity SC5+ . More inert in the syngas will lower partial pressures of H2 and CO, thereby reducing SC5+ .

M.J.A. Tijmensen et al. / Biomass and Bioenergy 23 (2002) 129 – 152

133

Table 1 Di;erences between %xed bed and slurry FT synthesis processes

Fixed bed

Slurry

Maintenance and labour intensive and long down time due to periodical catalyst replacement Scale up is straightforward, by multiplying tubes, economies of scale however limited [6]

Little down time due to on-line catalyst replacement. Lower catalyst consumption [9,10]

Conversion e=ciency Once through conversion

Up to 80% is assumed to be possible

C5+ selectivity

¿ 90% possible, negatively in@uenced by inert

Pressure drop

3–7 bar

High average conversions (once through), up to about 80% [10,11]. ¿ 90% possible, negative in@uence of inert seems lower [8] ¡ 1 bar [9]

Technical aspects Wax=catalyst separation

Performed easily and at low costs

More di?cult for commercial application; although solutions are reported [10] Easy 1.5 –2 times higher, so more thorough cleaning required

Proven technology. Advanced reactors are likely to have higher once through conversion

Overall considered proven technology; wax=catalyst separation is complex

Economic O&M Economies of scale

Process control Sulphur poisoning Status

2.1.2. FT reactors There are three main kinds of FT reactors: the @uidised bed reactor, the %xed bed reactor and the slurry phase reactor. The %xed bed (Fig. 3) and the slurry reactor (Fig. 4) are the most promising according to many authors [e.g. 6], some favouring the slurry phase reactor [e.g. 4] and some favouring the %xed bed [7]. Speci%c biomass related advantages for either %xed bed or slurry cannot be given clearly, though sensitivity for inert (relevant for some biomass-derived syngas compositions) seems less in a slurry reactor. The main disadvantage of the slurry reactor is the need for catalyst=wax separation, on which no public information appears to be available. Table 1 summarises some key di;erences between the %xed bed and slurry phase reactors [12]. 2.1.3. Hydrocracking When diesel is the desired %nal product, the FT product requires hydrocracking. Hydrogen is added to remove double bonds, after which the FT-liquids are cracked catalytically with hydrogen. Depending on the wax cracking conditions, mainly diesel or kerosene is

Possible, but scale-up is di?cult and exact economies of scale not clearly reported [6,11]

Table 2 Typical product distribution for di;erent wax hydrocracking conditions, (in wt%). In addition a small percentage C3 =C4 -fraction is formed [8] Product split

Gasoil mode (%)

Kerosene mode (%)

Naphtha Kerosene Gasoil (diesel)

15 25 60

25 50 25

produced. The overall carbon e?ciency of the hydrocracking step is close to 100% [9]. Also, hydrocracking conditions can be altered relatively simple to obtain a desired product mix (Table 2). The FT products are totally free of sulphur, nitrogen, nickel, vanadium, asphaltenes, and aromatics which all are typically found in mineral oil products. FT-diesel, with a very high cetane number, 3 can also be used as a blendstock to improve the quality of 3

Cetane number is a primary measure of diesel fuel quality. It is essentially a measure of the delay before ignition. The shorter the delay the better—and the higher the cetane number.

134

M.J.A. Tijmensen et al. / Biomass and Bioenergy 23 (2002) 129 – 152

normal diesel. The FT naphtha has a much lower octane 4 number than ‘normal’ naphtha. FT-kerosene for aviation still needs approval of several product specs. Based on the current product speci%cations and requirements, these products could therefore have less value than ‘normal’ naphtha and kerosene. But like FT-diesel they contain no sulphur or other contaminants. Besides reducing emissions to air, FT-liquids are expectedly well suited for use in fuel cell vehicles (FCVs), which require very clean fuel to prevent damage to fuel cell catalyst. This is a very important characteristic for the somewhat longer term when FCVs start penetrating the market [13], as %rst generation FCVs are likely to onboard reform diesel or gasoline. The diesel markets may be the %rst application of FT-fuels however. 2.2. Biomass gasi>cation 2.2.1. Pre-treatment of feedstock A wide variety of biomass resources can be used as feedstock. Wood, agricultural wastes, organic wastes, and sludges are each potential fuels. However, in this study clean (poplar) wood is assumed to be used as feedstock. Clean wood gives a relatively clean syngas, with low levels of contaminants. On the longer term wood from dedicated plantations may become a major source of renewable biomass [see e.g. [14]]. Pre-treatment prior to gasi%cation is required and generally consists of screening, size reduction, magnetic separation, ‘wet’ storage, drying and ‘dry’ storage [15]. Moisture content of the ‘wet’ poplar chips delivered is assumed to be 30%. Drying is generally the most important pre-treatment operation, necessary for high cold gas e?ciency at gasi%cation [16]. Drying reduces the moisture content to 10 –15%. Drying can either be done with @ue gas or with steam. Since, as will be explained, signi%cant amounts of low-quality steam are generated in the FT process, steam drying is preferable (e.g. a Niro steam dryer [17]). Also, steam drying results in (very) low emissions and may

4 Octane number is a quality rating for gasoline, indicating the ability of the fuel to resist premature detonation and to burn evenly when exposed to heat and pressure in an internal combustion engine. Normal gasoline has a octane number of 87–89.

eventually be safer with respect to risks for dust explosion. 2.2.2. Gasi>cation Conversion of biomass to an H2 and CO containing feed gas that is suited for FT synthesis takes places through gasi%cation. Gasi%cation can take place at different pressures, either directly heated or indirectly heated (lower temperatures), and with oxygen or with air. Direct heating occurs by partial oxidation of the feedstock, while indirect heating occurs through a heat exchange mechanism. Since for economic and e?ciency considerations, the capacities investigated in this analysis start at 100 MWth , only CFB gasi%ers have been taken into consideration [1,2,18,19]. Some key advantages and disadvantages of each gasi%cation method are shown in Table 3. To cover the wide range of gasi%cation methods, di;erent gasi%ers currently available and=or under development were selected for further study (see Table 4). These gasi%ers produce a wide range of syngas compositions representing the reasonably maximum possible variation in CO : H2 ratios that can be obtained. The syngas produced by the di;erent gasi%ers contain various contaminants: particulates, condensable tars, alkali compounds, H2 S, HCl, NH3 , HCN and COS [23]. However, no full data sets of syngas compositions including all these contaminants are available for the gasi%ers considered. Therefore, some estimations and assumptions are required (Section 2.5). 2.3. Syngas processing The syngas, produced by the gasi%cation of biomass, consists mainly of H2 , CO, CO2 and CH4 . Their shares in the syngas can be tailored to the needs of the FT process by methane reforming (converts CH4 with steam to CO and H2 ), a shift reaction (adjusts the H2 =CO ratio by converting CO with steam to H2 and CO2 ) and CO2 removal, which reduces the amount of inert gases for the FT synthesis. For methane reforming, the autothermal reformer (ATR) has been selected. An amine treating process is used for CO2 removal. More extensive descriptions on component performance data are given in background reports [18,24,19].

M.J.A. Tijmensen et al. / Biomass and Bioenergy 23 (2002) 129 – 152

135

Table 3 Main technical aspects gasi%cation method

Pressurised + Pressurised downstream equipment is smaller and generally more economical at larger scales (see Section 4.1) − Higher costs of gasi%er at small scale − High risk in keeping constant mass @ow in gasi%er, operational experience so far limited to demonstration projects [20] Oxygen

− Air separation plant needed, especially at small scales relatively

expensive + No dilution of syngas by N2

Direct + Less tars produced; the presence of tars in the syngas is one of the biggest problems when gasifying biomass

Atmospheric − Larger downstream equipment needed + Less costs at small scale (see Section 4.1) + Signi%cant commercial experience with airblown direct systems Air + Cheaper −N2 diluted syngas, has negative in@uence on C5+

selectivity − Larger equipment needed downstream Indirect + No N2 dilution even if air is used − Bigger tar problem − ¿ Currently demonstrated (BCL)

2.4. Power generation with a combined cycle

2.5. Gas cleaning

It is possible to use the o;gas from the FT reactor in a gas turbine (combined cycle) for additional electricity production. At high pressure, the o;gas is mixed with pressurised air and combusted at about ◦ 1100 –1300 C. Expansion of the resulting hot @ue gas generates power. Part of the power is used to drive the air compressor. In case of BIG-FT systems, the caloric value of the o;gas may be too low for (direct) combustion in a gas turbine. Typically, reasonable minimum heating values for commercial gas turbines are 4 –6 MJ=Nm3 , assuming some modi%cations on burners and fuel manifolds [15,25]. Co-%ring of natural gas avoids problems by raising the heating value of the gas, and increases the thermal e?ciency due to the larger scales of the turbine that are possible with co-%ring compared to o;gas utilisation alone. Co-%ring is therefore included in this analysis. Various (gas) streams in the whole process require cooling, e.g. the exhaust gas from the gas turbine and the syngas after gasi%cation. Superheated steam can be generated at these places and expanded in a (partly) condensing steam turbine to generate electricity. Low temperature steam can also be used: for drying and other steam demanding processes such as the shift reactor.

2.5.1. General The syngas produced by the gasi%cation process contains di;erent kinds of contaminants, viz. particulates, condensable tars, alkali compounds, H2 S, HCl, NH3 and HCN [23]. These contaminants can lower activity in the FT synthesis due to catalyst poisoning. Sulphur is an irreversible poison for the cobalt and iron catalysts (and to a smaller extent for the shift and reformer catalysts), because it will stick to active site. Tolerance for contaminants is low and ‘deep’ cleaning is required. Two distinct routes of cleaning will be considered in this study: ‘wet’ low temperature cleaning and ‘dry’ high temperature cleaning. 2.5.2. Conventional ‘wet’ low temperature cleaning Conventional ‘wet’ low temperature cleaning (Fig. 5), as described in [15,23], is being proposed applied to clean the fuel gas for BIG=CC installations. However, cleaning requirements for the FT synthesis are much more stringent than for BIG=CC systems. Speci%cations for FT are given in Table 5 and compared to syngas compositions typical for CFB gasi%cation of clean wood. Since at present, relatively clean natural gas is the common feedstock for FT synthesis, actual cleaning speci%cations are not known

136

Table 4 Operating characteristics of the gasi%ers evaluated in this paper, based on poplar wood

BCL (Batelle Columbus gasi%er) [1]

IGT (Institute of Gas Technology) [18]

IGT+a (IGT with process adjustment, based on estimates) [21]

EP (Enviro Power with dolomite tar cracker) [15]

TPS Termiska Processer with dolomite tar cracker) [19]

Gasi%er process type

Indirect, airblown, atmospheric

Direct, oxygen blown, pressurised

Direct, airblown, pressurised

Direct, airblown, atmospheric

Characteristics P (bar) T (K), exit Moisture dry biomass Pilot size (dry tons=day) Flowrate dolomite (kg=kg wet)c Flowrate air=oxygend Steam (kg=kg wet input) Yield (kmol=dry tonne) LHV syngas (MJ=Nm3 wet gas) Gasi%er e?ciencye H2 =CO ratio

2b 1136 10% 200 —c 1:46 kg=kg dry 0.19 45.8 13.9 86.8 0.45

34 1255 15% 100 (Knight 1999) 0 0:3 kg=kg dry 0.34 82.0 7.3 80.7 1.39

20.3 1241 15% No pilot 0 0:3 kg=kg dry 0.6 123.1 4.8 80.9 2.0

22 1223 15% —c 0.0095 — 0.34b 113.3 5.8 88.6 0.73

1.3 1173 15% 270 0.0268 1:4 kg=kg wet 0.34b 112.1 5.2 80.0 0.77

Composition (mol% [dry]) H2 O H2 CO CO2 CH4 C2+ C2 H4 C 2 H6 N2 + Ar Others

19.9 [0] 16.7 [20.8] 37.1 [46.3] 8.9 [11.1] 12.6 [15.7] 4.8 [6.0] 4.2 [5.2] 0.6 [0.74] 0 ¡ 0:3

31.8 [0] 20.8 [30.5] 15.0 [22.0] 23.9 [35.0] 8.2 [12.0] 0.3 [0.5]

50.6 [0] 15.68 [31.7] 7.83 [15.85] 17.71 [35.9] 5.73 [11.6] 0

13.55b [0] 10.03 [11.6] 13.83 [16.0] 15.4 [17.8] 7.26 [8.4] 0.48 [0.62]

0.40 [0.8] ¡ 0:3

38.9 [45.3] ¡ 0:3

13.55 [0] 13.25 [15.3] 17.22 [19.9] 12.22 [14.1] 2.82 [3.26] 0.96 [1.11] 0.94 0.02 39.20 [45.3] ¡ 0:3

a Two

gas compositions are included for pressurized, oxygen blown gasi%cation: one is at ‘standard’ conditions (IGT), the other at operating conditions leading to maximized hydrogen production, which is obtained by increased steam addition and reduced pressure. This is a more theoretical case, but allows for exploring the consequences for having a more optimal CO:H2 ratio at the cost of higher steam consumption, lower pressure and a lower heating value of the (wet) syngas. b Assumption. c Not available. d Oxygen of 99.5% purity is used; production is assumed to require 305 kWh per tonne oxygen for 95% purity. This is somehow scale-dependent [22]. Oxygen of 99.5% purity will require 15% extra energy [22]. e Gasi%er e?ciency is de%ned as [energy content syngas=energy content biomass input], based on LHV. Energy content of steam and air=oxygen added is not taken into account.

M.J.A. Tijmensen et al. / Biomass and Bioenergy 23 (2002) 129 – 152

Gasi%er name:

M.J.A. Tijmensen et al. / Biomass and Bioenergy 23 (2002) 129 – 152

Tar cracker

Cyclone separator

Bag filter (optional: second bag filter)

COS hydrolisation (not if Amine treating is used)

Scrubber (water + NaOH)

137

Scrubber (with H2SO4)

ZnO guard bed

Fig. 5. Schematic view of ‘wet’ low temperature cleaning. Table 5 Contaminant concentrations (wt%) and their maximum values for FT synthesis (ppb)

Contaminant present in dry feedstock (in syngas)

poplar wood adapted from [18]

Assumed cleaning requirements FT in ppb

Cleaning e?ciency required

Required cleaning steps (‘wet’ low temperature cleaning)

Results += requirement achieved ?= some uncertainty

Ash (particulates)

1.33

¿ 99:9%

N (HCN + NH3 )

0.47

¿ 99:9%

S (H2 S + COS)

0.01

¿ 99:9%

Alkalis

0.1a

¿99.9%

Cl (HCl)

0.1

¿ 99:9%

Pb and Cu

Not known

—

Tars

—b

¿ 99:9%

Cyclone separator, bag %lters= scrubber Scrubber (possibly with H2 SO4 ), Sul%nol D also removes HCN and NH3 Scrubber, possibly COS hydrolisation unit or Sul%nol D necessary, ZnO guard bed During cooling down alkalis condense on particulates, possibly also on vessels (and thereby polluting them) Absorbed by dolomite in tar cracker (if used), reaction with particulates in bag %lter, scrubber (possibly with NaOH) Condense on particulates, but actual behaviour has not been studied Condense on particulates and vessels (and thereby polluting them) when syngas is cooled below ◦ 500 C

+? ++, ZnO guard beds are also used for natural gas based FT +?

? potential tar problem, limited experience with complete removal or conversion for biomass

a Adapted b Not

from [16] for miscanthus; value would be lower for clean wood. known, but order of magnitude is g=Nm3 .

for some speci%c biomass contaminants. Therefore, some speci%cations are estimates. The speci%cation for sulphur, however, is explicitly known, since it is also present in natural gas and known to irreversibly poison the FT catalyst. Although with proper sizing and maybe addition of active coal %lters it is likely that the speci%cations can be met, this needs to be tested and veri%ed in practice. In particular potential problems with

tars give rise to discussion and further research is required. 2.5.3. Advanced ‘dry’ hot gas cleaning Hot gas cleaning consists of several %lters and separation units in which the high temperature of the syngas can (partly) be maintained, potentially resulting in e?ciency bene%ts and lower operational costs. Hot gas cleaning is speci%cally advantageous when preceding

138

M.J.A. Tijmensen et al. / Biomass and Bioenergy 23 (2002) 129 – 152

Table 6 Possible choices in system components leading to di;erent system con%gurations for BIG-FT systems

Gasi%er

Gas cleaning

Reforming

Shift

CO2 removal

FT system

FT data

BCL IGT IGT+ EP TPS

Low temp. High temp.

ATR No ATR

Intern (iron) Extern (cobalt)

Yes No

Full conversion once through Fixed bed slurry

Once through e?ciency

a reformer or shift reactor, because these process steps have high inlet temperatures. When FT synthesis is applied directly after the gas cleaning, the syngas has ◦ to be cooled to 200 C anyway, and the potential bene%ts are expectedly less apparent. Hot gas cleaning after atmospheric gasi%cation does not improve e?ciency, because the subsequent essential compression requires syngas cooling anyway. Hot gas cleaning is not a commercial process yet; some unit operations are still in the experimental phase. Disadvantageous for the application of hot gas cleaning in this case is the high puri%cation requirements of FT synthesis (see Table 7). It is uncertain if hot gas cleaning can meet these standards in a foreseeable timeframe. Research so far is mainly focussed on developing hot gas cleaning to meet the requirements for BIG=CC installations, for which fuel gas requirements are less severe. There are no commercial processes for the high temperature removal of nitrogen compounds, halides, alkali metals and heavy metals yet, although various solutions are worked on [15,23,27]. 2.5.4. Discussion Conventional ‘wet’ low-temperature syngas cleaning is the preferred technology in the short term [23]. This technology will have some e?ciency penalties though and requires additional waste-water treatment but there is little uncertainty at present about the cleaning e;ectiveness of such systems with respect to IG-CC installations, both for coal and biomass-%red facilities [25]. However, actual testing with biomass=FT systems is still necessary to ensure the e;ectiveness for these systems. Within 10 years hot gas cleaning may become commercially available for BIG-CC installations [15,28], but opinions di;er. Even when BIG-CC requirements are met, signi%cant improvements are necessary to

meet the much more severe FT requirements. As a result, hot gas cleaning is considered as an advanced option for the longer term in this analysis. To show the possible e;ect of hot gas cleaning on the performance of BIG-FT systems, it is included in the system modelling work. 2.6. Potential system con>gurations and selection 2.6.1. Key elements In theory, a large number of system con%gurations to convert biomass to FT-liquids and power is possible (Table 6). We make a %rst distinction between two main categories %rst: 1. Full conversion FT with the possible use of a gas turbine, aimed at maximised FT-liquids production. 2. Once through FT, with co-%ring of the o;gas with natural gas in a gas turbine. For all concepts an external shift reactor will be used, implying the use of a cobalt catalyst for FT. Since it is di?cult to predict a speci%c SC5+ for biomass-derived syngas, e.g. due to di;erent inert percentages, a realistic SC5+ range of 73.7–91.9% is modelled. This corresponds to chain growth probabilities of 0.8– 0.9. Higher C5+ selectivity seems unlikely to be obtainable for biomass syngas. As a conventional option a CO conversion of 40% per pass is assumed for the full conversion concepts. To increase the overall CO conversion, gas recycle has been employed. As an advanced option 60% and 80% once through CO conversion is assumed for the once through concepts. Advanced FT reactors, either %xed bed or slurry are expected to obtain higher once through CO conversion [8]. ◦ After the FT synthesis cooling to 50 C makes it possible for the C5+ fraction to be separated as liquid.

M.J.A. Tijmensen et al. / Biomass and Bioenergy 23 (2002) 129 – 152

Further cooling to separate C3 and C4 could be done, but whether this economically attractive is questionable. In the concepts modelled, the C1 –C4 fraction is assumed to be used in a gas turbine (combined cycle) for power generation. 2.6.2. Full conversion FT, with possible use of a gas turbine The goal of the full conversion concepts is to maximise the yield of FT-liquids (and to a smaller extent on power production). Systems modelled both include and exclude reforming, to study the pay o; between higher FT yield and lower capital investments. The IGT plus option is modelled without reformer and shift, otherwise the basic idea of creating a syngas with a ‘perfect’ H2 =CO ratio would not make sense. In case of the direct atmospheric gasi%cation systems, reforming with air will lower the H2 and CO content in the syngas because some H2 and CO have to be burned to obtain the necessary process heat. Therefore, this con%guration is not modelled. Amine-based CO2 removal was included in all con%gurations, as less CO2 allows for a high C5+ selectivity and will therefore increase the amount of FT-liquids produced. But on the other hand, CO2 removal is an expensive process and it is questionable if the high costs are justi%ed by the increased production of FT-liquids. The impact of in- and

139

excluding CO2 removal is therefore investigated further. 2.6.3. Once through FT, with co->ring in a (150 MWe) natural gas-based gas turbine The once-through concepts produce both FT-liquids and power. All gasi%ers are modelled without ATR and without CO2 removal, so FT synthesis is not maximised. This may result in lower capital costs and possibly somewhat higher overall energy e?ciencies. The IGT gasi%ers, in turn, are also modelled with an amine treating process, because of their high CO2 content of their syngas. 3. System calculations 3.1. Modelling and results Modelling of the various concepts has been performed with the @owsheeting program Aspenplus . A basic @owsheet is presented in Fig. 6. The gasi%cation processes have however not been modelled in Aspen, because the fuel gas compositions resulting from most biomass gasi%cation processes are determined by kinetics instead of equilibrium conditions and therefore very hard to model. Thus, fuel gas compositions from literature are used as starting point for the calculations. The base capacity for all system calculations

Fig. 6. Base Aspen Plus @owsheet used for the calculation of the mass and energy balances.

140

M.J.A. Tijmensen et al. / Biomass and Bioenergy 23 (2002) 129 – 152

Table 7 Input data for Aspen modelling. Pressures given in bar

Dryer: per tonne biomass 0:41 tonne steam needed for drying to 10% moisture content, 0.33 for drying to 15%. ◦ ◦ ‘Wet’ cold gas cleaning: Tinlet = 400 C, Toutlet = 40 C below dew point, Rp = −0:5 at p ¿ 30, Rp = −0:3 at 10 ¡ p ¡ 30, Rp = −0:2 at p ¡ 10 ◦

Hot gas cleaning: Toutlet = 450 C, Rp = −1 ◦

◦

ATR: p = 33–44, Rp = −0:5, Tinlet = 400–450 C, Toutlet = 950 C, Tair = 600 C, Toxygen = 300 C ◦

Shift: 15 ¡ p ¡ 70, Rp = −0:5, Tinlet = 330 C ◦

◦

FT reactor: p = 40, Rp = 5, Tinlet = 200 C, Toutlet = 240 C,e;ective H2 =COoutlet ¿ 0:4, recycle rate (mol@ow recycle=mol@ow ◦ fresh) ¡ 2, FT steam at p = 22 and T = 230 C. ◦

◦

Gas turbine: T = 1200 C but if LHV ¡ 6 MJ=Nm3 than T = 1100 C, expander: isentropic e?ciency = 0:89 (0.9 co-%red) mechanical = 1, compressor: isentropic e?ciency = 0:91 mechanical = 1, pout = 1:2 (0.2 needed for heat exchanger), T@ue gas (after ◦ heat exchanger) ¿ 170 Ca ◦

◦

Steam turbine: pressure=temperature combinations (T in C): 70=500, 41.4=440, 22=375 (230 C for FT steam), steam to dryer: p = 12, steam to gasi%er: p = 1–34, steam for shift: p = 15:5–43, steam for ATR: p = 20–44, outlet pressure of steam turbine = 0:04 Heat exchangers: Rp = −0:5 at p ¿ 30, Rp = = − 0:3 at 10 ¡ p ¡ 30, Rp = −0:2 at p ¡ 10, maximum syngas. Syngas heating ◦ is 400 Cb a Taken b Due

from [15]. Minimum outlet temperature is due to environmental restraints. to coking problems at higher temperatures [26].

Table 8 Overall energy e?ciencies (LHV) of the full conversion (40% per pass) concepts and net power (expressed in MWe ) and FT-liquids output (expressed in MWth ). = 0:8 corresponds with SC5+ = 73:7%; 0.85 with 83.5%; 0.9 with 91.1%

Concepta

BCL R

= 0:80 e;% FT-liquids (MWth ) Power (MWe ) = 0:85 e;% FT-liquids (MWth ) Power (MWe ) = 0:90 e;% FT-liquids (MWth ) Power (MWe )

45.1 123.9 41.4 47.0 139.1 33.3 48.0 154.8 21.3

BCLR R-nt 30.1 123.9 −13:6 33.5 139.1 −16:3 38.2 154.8 −14:6

BCL

IGT R

IGT R-hg

IGT

IGT+

EP R

35.9 66.8 65.0 37.1 75.9 60.4 38.1 83.4 56.5

44.7 129.9 34.2 47.7 150.7 24.4 50.1 168.7 15.2

46.1 132.6 36.6 49.1 153.4 26.8 51.5 171.4 17.4

46.0 81.1 87.5 47.4 91.9 82.2 48.2 101.7 75.1

44.6 78.4 85.4 44.9 89.8 75.0 47.3 97.9 75.7

41.6 100 52.5 44.8 113.3 50.9 45.4 124.6 42.0

EP R-nt 25.4 100 −6:7 29.4 113.3 −5:5 32.2 124.6 −6:5

TPS

42.4 64.6 90.9 43.4 73.2 86.2 44.5 80.5 82.9

32.9 83.1 37.6 34.5 94.2 32.5 35.8 103.6 27.6

a Explanation of used codes: BCL, IGT, IGT+, EP and TPS = gasi%er names, R = reformer used, nt = no gas turbine, hg = hot gas cleaning.

is %xed at 367 MWth LHVwet (80 tph biomass d.b., at 30% moisture). Table 7 summarises the key data used for Aspen modelling. The overall energy e?ciencies 5 of the full conversion concepts are presented in Table 8. Here, a distinction is made between outputs by means of FT-liquids and by net power production (or use). 5 The overall energy e?ciency of the systems is de%ned as: sum of all outputs=total biomass input.

The BCL-R and the EP-R concepts produce o;gas with a caloric value probably too low for (direct) use in a gas turbine, as part of the methane and ethane is converted to FT-liquids. The gas turbine is of particular importance for those concepts because without the gas turbine, a substantial amount of energy, in the form of light hydrocarbons (C1 –C4 ) or compressed N2 , would be wasted. Co-%ring with natural gas could be necessary for these concepts in any case, to upgrade the heating value of the o;gas.

M.J.A. Tijmensen et al. / Biomass and Bioenergy 23 (2002) 129 – 152

141

Table 9 Overall energy e?ciencies (LHV) of the 60% conversion once through concepts and net power (expressed in MWe ) and FT-liquids output (expressed in MWth ). = 0:8 corresponds with SC5+ = 73:7%; 0.85 with 83.5%; 0.9 with 91.1%

concepta

BCL

IGT

IGT-A

IGT+

IGT+-A

TPS

= 0:80 e;% FT-liquids (MWth ) Power (MWe ) = 0:85 e;% FT-liquids (MWth ) Power (MWe ) = 0:90 e;% FT-liquids (MWth ) Power (MWe )

35.8 49.9 81.3 37.2 56.6 79.8 38.0 62.2 77.2

44.1 58.9 103.0 44.6 66.8 97.0 45.6 73.5 93.7

43.9 58.9 102.2 44.5 66.8 96.4 45.4 73.5 93.1

43.5 57.8 101.9 44.6 65.4 98.1 45.4 72.0 95.0

43.6 57.8 102.4 44.7 65.4 98.8 45.7 72.0 95.6

43.6 54.8 105.1 44.6 62.0 101.8 45.5 72.0 98.6

33.9 69.3 55.2 35.2 78.5 50.7 36.3 86.3 47.0

a Explanation

of used codes: A = Amine treating used.

Table 10 Overall energy e?ciencies (LHV) of the 80% conversion once through concepts and net power (expressed in MWe ) and FT-liquids output (expressed in MWth ). = 0:8 corresponds with SC5+ = 73:7%; 0.85 with 83.5%; 0.9 with 91.1%

Concept

BCL

IGT

IGT-A

IGT+

IGT+-A

TPS

= 0:80 e;% FT-liquids (MWth ) Power (MWe ) = 0:85 e;% FT-liquids (MWth ) Power (MWe ) = 0:85 e;% FT-liquids (MWth ) Power (MWe )

37.7 66.6 71.9 38.9 75.4 67.4 40.4 82.9 65.3

46.4 78.6 91.5 47.8 89.1 86.3 49.0 98.0 81.9

46.2 78.6 91.0 47.7 89.1 85.9 48.9 98.0 81.6

45.6 77.0 90.4 47.0 87.3 85.3 48.2 96.0 81.0

45.7 77.0 90.8 47.1 87.3 85.8 48.3 96.0 81.4

47.2 73.0 100.4 48.6 82.7 95.7 49.8 91.0 91.6

36.5 92.4 41.5 38.2 104.6 35.6 39.7 115.1 30.5

The overall energy e?ciencies of the once through concepts are presented in Tables 9 and 10. All once through concepts make use of a 150 MWe gas turbine, so that a certain degree of natural gas co-%ring is necessary. The net power output in the table is therefore allocated to the fraction of the energy input of biomass-derived o;gas. 3.2. Discussion of results The modelled concepts give the following insights: 1. The concepts with high overall energy e?ciency are based on pressurised gasi%ers. When high C5+ selectivity (91.9%) is assumed, IGT-R has an LHV e?ciency of 50.1%, when hot gas cleaning is used 51.5%. The 80% once through concepts show that e?ciencies of near 50% are obtainable for the EP and the IGT gasi%ers, even without the use of a reformer.

2. CO2 removal has little e;ect on overall energy e?ciency—but does have an e;ect on the amount of inert and thus on liquid selectivity. Question remains if the investment is justi%ed by the improvement of selectivity. 3. Higher C5+ selectivity leads to higher overall ef%ciency for all concepts. The obtainable SC5+ is uncertain, but will be in the given range (73.7– 91.9%). When much inert is present (as is the case with air gasi%cation), SC5+ will probably end up in the lower part of the used range. When little inert is present SC5+ will be higher. 4. The concepts with 80% conversion have higher ef%ciencies than the concepts with 60% conversion. So a high overall CO conversion has a bene%cial e;ect on e?ciency. 5. Pressurised concepts have higher overall e?ciencies than atmospheric concepts. The di;erence is about 10% points. This is mainly due to the high electricity consumption of syngas compression

142

M.J.A. Tijmensen et al. / Biomass and Bioenergy 23 (2002) 129 – 152

when direct, air-blown, atmospheric gasi%ers are used. 6. Methane reforming concepts are most sensitive to C5+ selectivity and only improve e?ciencies when SC5+ is high. For the BCL concepts methane reforming does even improve e?ciency at low SC5+ , due to the high C2+ content of the syngas. 7. Hot gas cleaning improves the e?ciency for pressurised concepts, with an average of 1–2% points.

4. Economics

A maximum size for each unit is taken into account, above which increasing scale is no longer (economically) attractive (see Table 11). When the total capacity of a conversion unit exceeds this maximum component capacity, cost %gures are composed by assuming that multiple units are built to meet the desired capacity. Therefore, overall scale factors are used for making cost estimates for much larger and smaller scales compared to the base capacity of 367 MWth . The gasi%ers used have di;erent maximum sizes. The maximum size of a gasi%er is mainly determined by two factors: whether the gasi%er operates at elevated pressure and whether the plant is located near a harbour. If road transport is considered, the dimensions of the road are of importance. When transport of the gasi%er to location can be done over water, much larger single units can be installed. For the TPS, BCL, IGT and EP gasi%ers, the maximum sizes assumed are 122, 200, 400 and 400 MWth HHV, respectively [21,30,13]. Exact maximum scales cannot be given because existing plants of such size are not yet realised in practice, but it can be expected that larger single pressurised gasi%cation reactors can be built than assumed here. This is an aspect that requires further research to enable more exact projections.

4.1. Basic principles

4.2. Calculation of production costs of FT-liquids

The costs for each system con%guration are based on cost data on component level, which were obtained from literature, vendor quotes and personal communication with experts. For components also present in BIG-CC installations, the free on board (FOB) price is multiplied with speci%c percentages to obtain the installed costs, see Table 11. For some components and installations, it is assumed, as a rule of thumb, that the FOB price should be tripled to obtain the installed costs (see Table 11). The capacity or scale of each component is derived from the energy and mass balances obtained through the Aspen modelling. The speci%c costs of most system components are a;ected by their capacity. The general relation

4.2.1. Investment costs The calculation of the overall total investment costs is done on basis of the cost data as presented in Table 11. Straightforward discounting is applied by annuity (interest rate of 10% and depreciation period of 15 years). Table 12 presents the total calculated investment costs for the various concepts. As an example, the breakdown of the capital costs for the IGT pressurised gasi%cation concepts is shown in Fig. 7. The pre-treatment, gasi%cation with oxygen and gas cleaning sections account for almost 75% of total capital costs. The use of an amine treating process for CO2 removal will add more than 10% to total capital costs.

Costsb =Costsa = (Sizeb =Sizea )R

4.2.2. Production costs of FT-liquids The calculated energy e?ciency and overall total investment costs are used to calculate the production costs of pure FT-liquids %rst (without hydrocracking). Key assumptions regarding interest rate and variable

The main energy losses in the model are caused by the gasi%cation section (thermal e?ciency ∼ 80%), the shifting and reforming section ( th ∼ 90%), the FT section ( th ∼ 78% for the converted CO and H2 ) and the combined cycle ( th ∼ 50%). It is important to note that the small hydrocarbons (C1 –C4 ) formed in the FT synthesis and used in the gas turbine, lead to energy losses both in the FT section and in the combined cycle. Also, from the FT section (35 bar outlet pressure) to the gas turbine (14 –16 bar inlet) the syngas loses pressure without producing electricity. An expansion turbine could possible be used to reduce this energy loss, but is not included here.

applies, where R is the scaling factor. For most system components used here, the value for R usually lies between 0.6 and 0.8.

M.J.A. Tijmensen et al. / Biomass and Bioenergy 23 (2002) 129 – 152

costs are presented in Table 13. The breakeven costs assume a power price of 0.057 US$=kWh (including a premium for green electricity applicable in the Dutch context). Electricity from the grid costs 0.03 US$=kWh. Production costs of FT-liquids in full conversion concepts, are presented in Fig. 8. Production costs for the 60% and 80% conversion once through concepts are shown in Figs. 9 and 10. The costs vary for the different concepts between 13 and 30 US$=GJ. For both concept categories (once through and full conversion), the pressurised (IGT) gasi%er, seems to turn out best. Pressurised systems have a big advantage over atmospheric systems. Also, the use of a gas turbine reduces production costs. To calculate costs including hydrocracking, a cracking unit is assumed to have installed costs of MUS$ 8.1 per 2000 bbl=day. Using 13.1% annual depreciation this comes down to 0.26 US$=GJ FT-liquid. The e?ciency of the hydrocracking process is assumed to be 98% and the H2 consumption is 176 g=GJ FT-liquid (0.17 US$=GJ FT-liquid). Overall costs for hydrocracking become 0.72 US$=GJ, being about 5% on top of the FT-liquids production costs. CO2 removal has a strong impact on overall production costs. For this reason production costs are also calculated without the amine treating process, for the same e?ciencies and selectivities. The caloric value of the o;gas, however, will be lower due to the higher CO2 content. This could hinder direct use of the o;gas in the gas turbine, but co-%ring is still possible. The concepts cannot be compared directly. Depending on the amount of inert in the FT reactor, C5+ selectivity will di;er. For example, the EP once through concept seems slightly better than the IGT once through concept. But the IGT process, with little inert in its syngas, is more likely to have high liquid selectivity than the EP process, which has a large amount of inert in its syngas. From Fig. 8 it appears that CO2 removal is a bad option for all full conversion concepts. However, if the reduction of inert by means of CO2 removal results in a liquid selectivity rising from 73.7 to 91:9 ( = 0:8 → 0:9), it does have a positive e;ect on production costs. The quantitative impact of the amount of inert on SC5+ is not known, however, and needs further detailed study and testing.

143

It is more e?cient to produce FT-liquids (78% thermal e?ciency) than power (approximately 55% ef%ciency using state-of-the-art combined cycles). If only C5+ is separated as liquid, it becomes more e?cient to produce power when the C5+ selectivity drops below 35%. When comparing once through with full conversion concepts one should realise that the once through concepts make use of more advanced FT reactors, with higher conversion per pass. The once through options also make use of a co-%red gas turbine. The co-%red gas turbine has higher energy e?ciency and bene%ts from lower speci%c investment costs. 4.3. Sensitivity analysis and longer term perspectives 4.3.1. Sensitivity analysis of key parameters The parameters that a;ect the %nal FT fuel costs strongest are shown in Table 14, including the potential range between which these parameters may vary. A sensitivity analysis has been performed for these parameters over the given range. Fig. 11 shows the results of doing so for the IGT-R concept at SC5+ = 83:5% (in which case 14% of the total energy output is power). Overall total investment costs for this concept are about 380 MUS$ and O&M is about 14 MUS$ annually. When power output is high (54% of the total energy output for the EP full conversion concept at SC5+ = 83:5%), the sensitivity to the electricity value is much stronger. When all parameters are set at their ‘base’ value, production costs of FT-liquids are 14 US$=GJ. This value can be compared to the current production costs of around 5 US$=GJ for diesel. Economies of scale have a considerable in@uence on overall production costs. When scale is between 100 and 400 MWth the overall scaling factor for the entire plant (with respect to overall total investment costs) is approximately 0.74. When capacities go beyond 400 MWth , the average scaling factor increases to 0.91 (indicating decreasing cost reductions). The results for doing so are shown in Fig. 12 for the IGT-R concept. Based on the component data and assumptions made, it can be concluded that scale e;ects level o; at very large capacities. Scale up reduces costs from 14 US$=GJ at 400 MWth to 12 US$=GJ at 1600 MWth, a reduction of 14%. When the capacity is smaller

144

M.J.A. Tijmensen et al. / Biomass and Bioenergy 23 (2002) 129 â&#x20AC;&#x201C; 152

Table 11 Basic costs for all units used with their maximum size (base costs are in relation to base scales). Costs expressed in MUS$

Base cost

Scale factor

Base scale

0.33 0.43 1.05 7.71 0.33 0.38

0.8 0.6 0.65 0.8 0.7 1

69:54 MWth 69:54 MWth 69:54 MWth 69:54 MWth 69:54 MWth 69:54 MWth

LHV LHV LHV LHV LHV LHV

367 367 367 367 367 367

3.24 13.0

0.7 0.7

69:54 MWth LHV 400 MWth HHV

0.7

400 MWth HHV

105 200 400 400

Gas cleaning Tar crackera Cyclonesa Gas coolinga Baghouse %ltera Condensing scrubbera Hot gas cleaningd

3.24 2.57 2.95 1.62 2.57 14.3

0.7 0.7 0.7 0.65 0.7

69:54 MWth LHV 69:54 MWth LHV 69:54 MWth LHV 69:54 MWth LHV 69:54 MWth LHV 400 MWth HHV

105 367 367 367 367 None

Compressors Compressore

12.0

0.85

13:2 MWe

None

Combined cycle Gas turbinea Modi%cations turbine LCV gasa HRSGa Steam turbine + condenser a Water + steam systema Coolinga

7.7 8% 3.38 4.48 0.43 0.95

0.7 0.7 0.8 0.7 0.9 0.3

25 MWe 69:54 MWth LHV 47:5 tonne=h 12:3 MWe 49:5 tonne=h 50:5 tonne=h

None None None None None None

Sub1: Total hardware costs Instrumentation and controlf Buildings Grid connections Site preparation Civil works Electronics Piping

5% of hardware 1.5% of hardware 5% of hardware 0.5% of hardware 10% of hardware 7% of hardware 4% of hardware

Sub2: Investment costs Engineeringg

15% of investment costs 0.75 0.7 0.6 0.65 1 1

24 tonne=h 400 MWth 2400 kmol=h 400 MWth 100 MW FT-liquid

None None None Input None None

Pre-treatment Conveyersa Grindinga Storagea Dryera Iron removala Feeding systema Gasi>ers Gasi%er TPSa Gasi%er BCL (incl. feeding)b Gasi%er EPc Gasi%er IGTc

Sub3: Total installed Oxygen planti ATR reactorj Shift reactork Amine treatingl FT reactorm ZnO bedsn a Costs

b Costs

costsh

23.0 30.3 0.45 con%dential 16.7 0.13

%gures based on %rst generation BIG=CC installations, taken from Faaij (1998). %gures taken from [29].

Unit maximum size (for scales considered)

M.J.A. Tijmensen et al. / Biomass and Bioenergy 23 (2002) 129 – 152

145

Table 12 Overall total investment costs in MUS$ for all concepts (367 MWth input; LHV wet (biomass supplied at 30% moisture content)

Full conversion concepts

BCL-R

BCLnt

BCL

IGT-R

IGT-hg

IGT

IGT+

EP-R

EP-R-nt

TPS

Total investment costs Idem without Sul%nol D

395 366

363 334

312 292

387 349

358 320

339 305

341 316

449 396

417 363

364 322

386 344

IGT 305

IGT-s 338

IGT+ 297

IGT+-s 310

EP 325

TPS 331

Once through concepts, 60% and 80% BCL Total investment costs 281

than 400 MWth , costs strongly increase, however, and small-scale production of FT-liquids is economically not feasible. 4.3.2. Short- and long-term perspective Results so far represent technology and performance that could be realised on shorter term. However, on the longer term various improvements may be feasible. These include: increasing C5+ selectivity, increasing scale, lowering feedstock costs, reductionof investment and O&M costs through technological learning, and application of hot gas cleaning. For the

short and long term the key assumptions made are summarised in Table 15. The type of assumptions for the short and longer term reasonably represent a %rst commercial plant and a plant that could be built after a period of 1–2 decades from now. The cost-breakdown for the short and long term is given in Fig. 13. Capital costs represent about 50% of the overall production costs of FT-liquids. Reduction of these capital costs for a third-generation plant, due to scaling up (12%) and technological learning (15%) therefore have a considerable impact on overall production costs. O&M costs may decrease almost

←−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−− c The Enviro Power gasi%er and the IGT gasi%er, both operating at elevated pressure, have likewise designs. Therefore, costs are assumed to be the same for both types of gasi%ers. Cost %gures for IGT taken from [29]. d These costs are an assumption. No real world data are available. e Vendor quote (Sulzor). f Percentages take from Faaij [16], valid for BIG=CC installations. Adding these percentages to capital costs (free on board, FOB) will result in installed costs. g Engineering is already included in the installed costs for the oxygen plant, ATR, shift, Amine treating, FT and ZnO units. h For the units below the percentages as discussed in footnote f are probably not valid. Therefore installed costs were calculated directly or the FOB price was tripled to obtain the installed costs. i 576 t=d oxygen production of 95%v purity has a capital cost of 31,000 US$ per t=d, for 1008 t=d the costs are 27,000 per t=d [22]. Oxygen of 99.5% purity requires 5% extra capital costs [22]. Costs are indexed to 1999 US$ using a Consumer Price Index of 0.816. j FOB price for the ATR is 10.1 million US$ [29]. Multiplying with three gives the installed costs. k Calculations were done on basis of Nm3 @ow and necessary reactor height. Given base unit gives a good representation of this. l Calculations were done on basis of Nm3 @ow and necessary reactor height. m Calculated for a %xed bed reactor. No cost data are available for slurry reactors. Main factor used is the amount of CO converted to FT-liquids (in MW, HHV based). One catalyst loading is included in these costs. n Assuming 1% wt of S (excluding N2 ) entering the ZnO bed, two guard beds of 3 m 3 are necessary. This will take about 2300 kilos of steel. Using a steel price of US $9.5=kilo, each guardbed will cost US $22,000 (FOB) or US $66,000 installed. To calculate overall total investment costs on top of the installed costs were increased with the following percentages [15]: Building interest. 1st yearo Building interest. 2nd year Project contingency

25% Installed costs × interest rate 75% Installed costs × interest rate 10% Installed costs × interest rate

Sub 4: Total investment costs Fees=overheads=pro%ts 10% of total investment Start-up costs 5% of total investment Overall total investment costs o Interest

rate used is 10%.

146

M.J.A. Tijmensen et al. / Biomass and Bioenergy 23 (2002) 129 â&#x20AC;&#x201C; 152

HRSG 11%

others 4%

pretreatment 21%

gas turbine 7% FT reactor 6% gasifier 18%

shift 1% cleaning section 18%

oxygen plant 15%

Fig. 7. Breakdown of installed costs for the IGT 80% conversion concept (367 MWth LHV).

proportionally with the reduction of the capital costs. The share of biomass feedstock costs (on themselves assumed to remain constant for the short and long term, see footnotes under Table 15) will decrease per GJ FT-liquid due to an increase of overall energy ef%ciency. Overall energy e?ciency will be higher for a third-generation plant, due to higher C5+ selectivity and higher (once through) CO conversion. As a result of the reduction in capital costs and biomass costs per GJ FT-liquid, production costs of FT-liquids could drop from over 14 to 9 US$=GJ.

It can be concluded that in the short term pressurised BIG-FT systems have production costs of FT-liquids higher than 14 US$=GJ. Atmospheric, air-blown gasi%cation-based systems result in much higher production costs. None of the concepts, either atmospheric or pressurised, have production costs that are competitive with current diesel costs of around 5 US$=GJ. By including a number of improvement options, production costs of FT-liquids can drop to around 9 US$=GJ. This is still above current diesel costs. Obviously, biomass-derived FT-liquids become more attractive with rising oil prices. Although projections for the future oil price development are highly uncertain, expected price ranges for diesel in 2020 go up from 5.5 to 7 US$=GJ [30]. Considering the inherent uncertainties in price estimates as composed from data reported in this study, the fact that not all improvement options for biomass-based FT synthesis are considered in this study, the longer term economic perspectives for biomass-derived hydrocarbons in the transport is not unattractive. In addition it must be stressed that biomass-derived FT-liquids have very di;erent characteristics than diesel from mineral oil. FT-liquids contain little or no contaminants (like sulphur and aromates) and are therefore suited for use in FCV when on board reforming (by means of partial oxidation) is applied. Such technology is already demonstrated at present, but requires very clean fuel to avoid

Table 13 Main assumptions for calculating the overall production costs of FT-liquids Cost factor

Unit

Input for 427 MWth

Annual Biomass costsb

13.1% of investment 2 US$=GJ (LHV)

10 560 000

Operational costs Maintenance Personnelc Dolomited Waste-water treatmente NaOH consumptione ZnO consumption [31] FT cat. consumption insurance

3% of investment 0.7 MUS$=100 MWth LHV 47.6 US$=tonne 0.21 MUS$l=75 MWth LHV 1.3 KUS$=tonne NaOH 33.3 KUS$=year 1% of annual depreciation

depreciationa

a real

interest rate = 10%; depreciation period = 15 years. GJ=tonne wet, 30% moisture, load factor = 8000 h=year. c [15] a scaling factor of 0.25 is used. d [15] dolomite is only used when by a tar cracker is applied. e [15] waste water treatment and NaOH consumption only if low temperature gas cleaning is applied. b 11:55

367 25 728 367 44 800 Con%dential

M.J.A. Tijmensen et al. / Biomass and Bioenergy 23 (2002) 129 – 152

147

30 alpha=0.8 alpha=0.85

US$/GJ FT liquid

alpha=0.9 ex-sulfinol

ex-sulfinol ex-sulfinol level

low premium high premium

0 BCL-R BCL-R-nt BCL

IGT-R IGT-R-h

IGT

IGT+

EP-R

EP-R-nt

TPS

Fig. 8. Production costs per GJ FT-liquid for the full conversion concepts (40% conversion per pass), assuming a power price of 0.057 US$=kWh. Scale used is 367 MWth LHV. = 0:8 corresponds with SC5+ = 73:7%; 0.85 with 83.5%; 0.9 with 91.1%. ex-CO2 represent concepts where CO2 removal is omitted.

alpha=0.8

alpha=0.85 alpha=0.9

15 10

level low premium

high premium

US$/GJ FT l iquid

US$/GJ FT liquid

alpha=0.8

alpha=0.85 15

alpha=0.9 level

low premium

high premium

0 BCL

IGT IGT-S IGT+ IGT+S

TPS

Fig. 9. Production costs per GJ FT-liquid for the 60% conversion once through concepts, assuming a power price of 0.057 US$=kWh. Scale used is 367 MWth LHV. = 0:8 corresponds with SC5+ = 73:7%; 0.85 with 83.5%; 0.9 with 91.1%.

poisoning of the fuel cell catalyst. On longer term, this can allow higher vehicle e?ciency compared to current diesel %red internal combustion engine vehicles [e.g. 2, 13]. In total, the real value of FT-liquids can than be considered to be higher than conventional diesel due to inherent higher e?ciency utilisation and (much) lower emission levels for e.g. sulphur, soot and other contaminants. The carbon neutral character of FT-liquids is of course the key di;erence with conventional diesel. Altogether, this may provide a basis for a premium on ‘green’ biomass-derived

IGT IGT-s IGT+ IGT+- EP s

TPS

Fig. 10. Production costs per GJ FT-liquid for the 80% conversion once through concepts, assuming a power price of 0.057 US$=kWh. Scale used is 367 MWth LHV. = 0:8 corresponds with SC5+ = 73:7%; 0.85 with 83.5%; 0.9 with 91.1%.

FT-diesel compared to conventional transport fuels (or kerosene). In the case of the Netherlands (but similar systems are observed throughout Western Europe), a premium is paid for green electricity varying between 0.024 and 0.038 US$=kWh (on top of the grid price of 0.028 US$=kWh). This is equivalent to a premium between 6.7 and 10.5 US$=GJ fuel on energy basis. Normal production costs of diesel are around 5 US$=GJ, depending on oil prices. Assuming that a similar premium is paid for ‘green’ FT-liquids as is done for ‘green’ electricity, ‘green’ diesel, naphtha and kerosene must

148

M.J.A. Tijmensen et al. / Biomass and Bioenergy 23 (2002) 129 – 152

the premium for green energy is lower than current the Dutch excise duty on diesel (as is the case for many other European countries). On the long term required ‘premiums’ could be far lower or even unnecessary to make biomass-derived FT-fuels cost competitive with diesel per kilometer driven.

Table 14 Main parameters used and ranges for the sensitivity analysis Parameter

Value

Range

Biomass costs Capital costs

US $2=GJ MUS $380 (varies per concept) US $0.057=kWh 8000 h=year 10 15 years

2.1–8.4 50 –175%

Electricity value Load factor Real interest rate Depreciation period

0.03– 0.07 7588–8760 6.25 –15 9.4 –22.5

5. Conclusions and recommendations This study presented a broad exploration of the possibilities to produce synthetic hydrocarbons from biomass via gasi%cation and Fischer Tropsch synthesis. A wide variety of potential conversion system con%gurations has been evaluated including energy e?ciencies and economics.

have production costs between 10.7 and 14.5 US$=GJ to be competitive. These values are also shown in Figs. 8–10. Such premium levels would make production of FT-liquids from biomass already economically attractive on the short term. It is interesting to note that

US$/GJ FT liquid

biomass costs capital costs

electricity value (low output) electricity value (high output) load factor

15 10 5 50%

real interest rate 75%

100%

125% 150% 175% 200%

parameter variation

depreciation period

Fig. 11. Sensitivity of production costs of the IGT-R concept to parameters used.

US$/GJ FT liquid

35 30 25 20 15 10 5

16 00

10 00

50 0

10 0

scale (MWth) Fig. 12. E;ect of scale on the production costs of FT-liquids; production costs of 14 US$=GJ are assumed at 400 MWth . Biomass feedstock costs are assumed constant here, where in practice biomass costs could slightly increase for larger scales due to higher logistic costs [32].

M.J.A. Tijmensen et al. / Biomass and Bioenergy 23 (2002) 129 – 152

149

Table 15 Assumptions for the short and long term

Short term (%rst commercial plant)

Long term (third generation)

IGT full conversion (40% once through, ex-Sul%nol) is the best concept Obtainable = 0:8 Scale of the system is 400 MWth Biomass costs are US $2=GJ

IGT once through 80% conversion (with high e?ciency gas turbine) is the best concepta Obtainable = 0:9 Scale of the system is 1600 MWth b Biomass costs are US $2=GJc Technological learning reduces capital costs with 15%d

a Hot gas cleaning has not been modelled for this concept since only a shift reaction is used for this concept. In that case still some cooling down is needed after the hot gas cleaning and consequently e?ciency advantage will be smaller. b An overall scaling factor of 0.91(with respect to overall total investment costs) is used. c In the longer term biomass costs may be lower, but larger scales will increase costs again [32]. d Technological learning can be assumed on longer term. This can be expressed by a progress curve; such curves are determined by a progress ratio. A progress ratio of x implies that each doubling of cumulative output leads to a (1 − x) × 100% reduction in costs. A progress ratio of 0.9 is used, applied for a third generation plant built. This results in 15% lower capital costs [e.g. 15].

US$/GJ

20 15

Biomass O&M

Investment co sts

5 0 short ter m

long term

Fig. 13. Cost breakdown for production of FT-liquids from biomass (excluding hydrocracking) for the short and the long term.

Systems applying pressurised gasi%ers (IGT and EP) have much better overall energy e=ciencies (42–50% LHV) than atmospheric systems (33– 40%). This is mainly due to the high electricity consumption of the syngas compressors when atmospheric gasi%ers are used. High CO conversion, either once through or after recycle of unconverted gas, and high C5+ selectivity are important for high overall energy e?ciencies. In the short term, production costs of FT diesel, naphtha and kerosene could be about 14 –16 US$=GJ. Capital costs represent about 50% of the overall production costs of FT-liquids. The pre-treatment, gasi%cation (with oxygen) and cold gas cleaning account for almost 75% of total capital costs. Biomass costs are 30% of total production costs (assuming a biomass price of 2 US$=GJ), and operation and maintenance about 20%.

In the longer term with large-scale production, high C5+ selectivity, high CO conversion and technological learning, production costs of FT-liquids could drop to 9 US$=GJ. Reduction of capital costs for a third-generation plant, due to scaling up and technological learning have a signi%cant impact on overall production costs. Feedstock costs per GJ FT-liquid decrease due to an increase of overall energy e?ciency, especially because of higher C5+ selectivity and higher (once through) CO conversion. When diesel is the desired %nal product (besides 60% diesel, 40% naphtha and kerosene are produced), the FT product requires additional hydrocracking. Hydrocracking will add about 5% to production costs. Production costs of ‘green’ FT-diesel, naphtha and kerosene are not competitive with conventional diesel prices, which cost about 0.14 US$=litre or 5 US$=GJ. In the longer term conventional diesel prices could go up due to higher oil prices, but still ‘green’ FT-liquids (9 US$=GJ) will not be competitive with expected diesel prices (5 –7 US$=GJ). There are several uncertainties with respect to the technology status. A very critical step in the whole system is gas cleaning. It still has to be proven whether the gas cleaning section is able to meet the strict cleaning requirements for FT synthesis, especially the high temperature concepts which are required to obtain the desired higher energy e?ciencies. Possibly amine-based CO2 removal is required for cleaning purposes, thereby raising production costs. The used syngas compositions as produced

150

M.J.A. Tijmensen et al. / Biomass and Bioenergy 23 (2002) 129 – 152

by the di;erent gasi%ers have strong in@uence on overall results. Most data are taken from literature and based on pilot-scale operating experience. The reliability of these data for large-scale gasi%ers is not known in great detail. Pressurised (oxygen) gasi%cation systems, having most promising economics and advantages of scale, still need further development. At present, only atmospheric air gasi%cation systems, operating at relatively small scale, have proved to be reliable. Not all possible concepts have been investigated though. Separating the C3 –C4 fraction as liquid product from the FT synthesis could be more advantageous than combustion of this fraction in a gas turbine. Also variable gasi%cation temperatures could make other reforming methods possible, like partial oxidation and steam reforming, due to a di;erent C2+ content. For the concepts modelled, however, reforming did not result in lower production costs. Lowering the pressure in the FT reactor will cause selectivity to drop, but on the other hand compression costs will also be lower. Using an iron catalyst could reduce production costs due to an internal shift reaction. This is to be investigated further. In the long term, the e?ciency of the concepts will be higher if high selectivity can be combined with high once through conversion. This could be realised in either %xed bed or slurry reactors. Costs for slurry reactors, which are not available yet, could be lower than for %xed bed reactors and are very likely to have better economies of scale. Heat integration of the total plant can also be improved. Power generation in the gas turbine will improve if the scale is larger and when on the longer term more e?cient turbines will enter the market [15]. This study did include co-%ring with natural gas for power generation with combined cycles as an ‘improvement’ option. Co-feeding FT synthesis was not included but could lead to some e?ciency and scaling bene%ts as well. The overall energy e?ciency, a critical parameter in obtaining good economic performance when more expensive cultivated biomass is proceeded, may be increased by optimised gas turbine technology [e.g. 15] and also improved selectivity for the FT process applied. With further catalyst and process development higher selectivities may be obtained, improving the net yield of the most valuable commodity of the process: transport fuel.

Technological learning over time and economies of scale were roughly included in the cost projections for longer term, but not investigated in great detail. Especially, for pressurised gasi%cation larger scales may prove to be more attractive than projected here and combinations with enriched air gasi%cation (eliminating the expensive oxygen production assumed in this study) may also reduce costs further. Altogether, the full technological improvement potential requires further study. Another key variable is the feedstock costs. (Cultivated) biomass is assumed to cost 2 US$=GJ. At present, this is low for Western Europe, but high compared to Brazilian biomass production costs (from Eucalyptus). Production of FT-liquids in regions where biomass feedstock is cheap (or by using co-products or biomass residues) will positively a;ect the economics of FT production further. It can be concluded that in the short term pressurised BIG-FT systems have production costs of FT-liquids higher than 14 US$=GJ. Atmospheric, air-blown gasi%cation-based systems result in much higher production costs. None of the concepts, either atmospheric or pressurised, have production costs competitive with current diesel costs of around 5 US$=GJ. By including a number of improvement options, production costs of FT-liquids can drop to around 9 US$=GJ. This is still above current diesel costs. Obviously, biomass-derived FT-liquids become more attractive with rising oil prices. Projections for the diesel price in 2020 range from 5.5 to 7 US$=GJ [13]. Considering the inherent uncertainties in price estimates as composed from data reported in this study, the fact that not all improvement options for biomass-based FT synthesis as discussed above are considered in this study, the longer term perspective for biomass-derived hydrocarbons for the transport sector is promising. In addition, it must be stressed that biomass-derived FT-liquids have very di;erent characteristics than diesel from mineral oil. FT-liquids contain little or no contaminants (like sulphur and aromates) and are therefore suited for use in fuel cell vehicles when on board reforming is applied. On longer term, this can allow a higher vehicle e?ciency compared to current diesel %red internal combustion engine vehicles [e.g. 2, 13]. In total, the real value of FT-liquids can than be considered to be higher than conventional

M.J.A. Tijmensen et al. / Biomass and Bioenergy 23 (2002) 129 – 152

diesel due to inherent higher e?ciency utilisation and (much) lower emission levels for e.g. sulphur, soot and other contaminants. Recommendations for further actions and research are: • The gas cleaning section needs special attention. Proper data sets of contaminants in the syngas must be made, with high detection accuracy. Deep, hot gas cleaning is promising as such, but will require even more development before su?cient cleaning is guaranteed. • Pressurised biomass gasi%cation must be developed for large-scale FT plants, but require demonstration at full scale. • For the use of biomass syngas in the FT synthesis, high liquid selectivity is desirable. The FT process (either a %xed bed or a slurry process) needs to be con%gured to ful%ll this need. • Development of sustainable forestry is necessary to ensure a large enough supply of clean wood. E;orts must be made to create a working biomass market and reduce prices of biomass, from various sources, over time. • Markets should be explored to determine what the outlets are for FT naphtha and kerosene, and if a ‘green’ fuels premium will be paid for these products, which is particularly relevant for the short term.

Acknowledgements The authors would like to thank all experts who assisted in providing comments, insights and information: Mr. Hogendoorn from Foster Wheeler, Rick Knight from the Institute of Gas Technology USA, Mr. Mitchell from IEA Coal Research UK, Kari Salo from Carbona Oy Finland, Mr. Vosloo from Sasol South Africa and various other experts consulted for this study. Shell Global Solutions, Amsterdam, the Netherlands, is thanked for supervising part of and contributing to the research activities carried out in this study. The Netherlands Energy Research Foundation ECN (Ren,e van Ree, Herman den Uil) is thanked for structural collaboration and exchange in this area. Samenwerkingsverband Duurzame Energie SDE (Professor

151

Kees Daey Ouwens) and Shell International (Peter Kwant) are thanked for co-funding part of the work. References [1] Williams RH, Larson ED, Katofsky RE, Chen J. Methanol and hydrogen from biomass for transportation. Princeton, New Jersey, USA: Princeton University=Center for energy and environmental studies, 1994. [2] Faaij A, Hamelinck C, Tijmensen M. Long term perspectives for production of fuels from biomass; integrated assessment and R&D priorities—preliminary results. In: Kyritsis S et al., editors. Proceedings of the First World Conference on Biomass for Energy and Industry. London, UK: James & James Ltd., 2001, vol. 1=2, p. 687–90. [3] Goldemberg, et al. The world energy assessment—energy and the challenge of sustainability. United Nations development programme. United Nations=economic and social a;airs. World Energy Council. New York, USA, 2000. [4] Larson ED, Jin H. Biomass conversion to Fischer-Tropsch liquids: preliminary energy balances. In: Overend R, Chornet E, editors. Proceedings of the Fourth Biomass Conference of the Americas. Kidlington, UK: Elsevier Science, 1999, vol. 1=2, p. 843–54. [5] Schulz H. Short history and present trends of FT synthesis. Applied Catalysis A: General 1999;186:1–16. [6] Agee MA. Studies in surface science. Catalysis 1998;119:931. [7] Sie ST, Krishna R. Fundamentals and selection of advanced FT-reactors. Applied Catalysis A: General 1999;186:55–70. [8] Espinoza RL, Steynberg AP. Low-temperature FischerTropsch synthesis from a Sasol perspective. Applied Catalysis A: General 1999;186:13–26. [9] Vosloo A. Fischer Tropsch expert at Sasol. Utrecht, The Netherlands: Colloquium at Utrecht University, 2000. [10] Van de Laan G. Kinetics, selectivity and scale up of the Fischer-Tropsch synthesis. PhD Thesis. University of Groningen, Groningen, The Netherlands, 1999. [11] Bechtel. Baseline design=economics for advanced FischerTropsch technology—quarterly reports. US Department of Energy. Pittsburgh, USA, 1991–1994. [12] Geerlings JJC, Wilson JH. Fischer-Tropsch technology— from active site to commercial process. Applied Catalysis A: General 1999;186:27–40. [13] Ogden JM, Steinbugler MM, Kreutz TG. A comparison of hydrogen methanol and gasoline as fuels for fuel cell vehicles: implications for vehicle design and infrastructure development Journal of Power Sources 1999;79:143–68. [14] Faaij A, Van den Broek R, Van Engelenburg B, Lysen E. Global availability of biomass for energy and possibilities and constraints for large scale international trade. In: Williams DJ, et al., editors. Proceedings of the Fifth International Conference on Greenhouse Gas Control Technologies. Collingwood, Australia: CSIRO, 2001 p. 1145 –51. [15] Faaij A, Van Ree R, Meuleman B. Long term perspectives of biomass integrated gasi%cation with combined cycle

152

[16]

[17]

[18] [19] [20] [21] [22] [23]

M.J.A. Tijmensen et al. / Biomass and Bioenergy 23 (2002) 129 – 152 technology—costs e?ciency and a comparison with combustion. Novem. Utrecht, The Netherlands, 1998. Faaij A, Van Ree R, Waldheim L, Olsson E, Oudhuis A, Van Wijk A, Daey Ouwens C, Turkenburg W. Gasi%cation of biomass wastes and residues for electricity production. Biomass and Bioenergy 1997;12(6):387–407. Pierik JTG, Curvers APWM. Logistics and pre-treatment of biomass fuels for gasi%cation and combustion. Petten, The Netherlands: Netherlands Energy Research Foundation ECN, 1995. Katofsky R. The production of @uid fuels from biomass. Princeton, NJ: USA: Princeton University=Center for Energy and Environmental Studies, 1993. Hamelinck C, Faaij A. Future prospects for production of methanol and hydrogen from biomass. Journal of Power Sources. 2002, in press. Lassing K, Olsson E, Waldheim L. TPS integrated gasi%cation combined cycle technology for waste processing. TPS Thermiska Processer AB, Nykoping, Sweden, 1995. Air products. Presentation on O2 separation techniques, 1991. Hogendoorn J. Project manager at Foster Wheeler. Written communication on biomass gasi%cation, 2000. Van Ree R, Oudhuis ABJ, Faaij A. Modelling of a biomass-integrated-gasi%er=combined-cycle (BIG-CC) system with the @owsheet simulation program Aspen-plus.

[24]

[25]

[26]

[27] [28] [29] [30] [31] [32]

Petten, The Netherlands: Netherlands Energy Research Foundation ECN, 1995. Tijmensen M. The production of Fischer-Tropsch liquids and power through biomass gasi%cation. Utrecht, The Netherlands: Utrecht University=Science Technology and Society, 2000. Consonni S, Larson ED. Biomass-gasi%er=aeroderivative gas turbine cycles. Part A: technologies and performance modelling. Part B: performance calculations and economic assessment. Cogen Turbo Power 1994. Portland, Oregon, USA, 1994. Williams RH, Larson ED, Katofsky RE, Chen J. Methanol and hydrogen from biomass for transportation, with comparisons to methanol and hydrogen from natural gas and coal. Princeton, NJ: Princeton University=Center for Energy and Environmental Studies, 1995. Mitchell SC. Hot gas cleanup of sulphur, nitrogen, minor and trace elements. London, UK: IEA Coal Research, 1998. Mitchell SC. Expert at IEA Coal Research. Written communication on advanced gas cleaning systems. 2000. Salo K. Director of Carbona Oy Finland. Written communication on biomass gasi%cation, 2000. Statistical Review of World Energy 1999. BP Amoco. 1999. www.shell.com. 1999. Jager B. Developments in Fischer-Tropsch technology. Dymmy 1997;107:219–24.

CAMPUS MONTERREY

APPENDIX 4 Alberto Mendoza, Porfirio Caballero, Juan A. Villarreal and Ricardo Viramontes. Performance of a Semi-Industrial Scale Gasification Process for the Destruction of Polychlorinated Biphenyls. J. Air & Waste Manage. Assoc. 56 (2006) 1599– 1606

Ave. Eugenio Garza Sada 2501 Sur Col. Tecnológico C.P. 64849 Monterrey, N.L., México Tel. 8358-2000 y 8358-1400

TECHNICAL PAPER

Performance of a Semi-Industrial Scale Gasification Process for the Destruction of Polychlorinated Biphenyls Alberto Mendoza Department of Chemical Engineering, Instituto Tecnolo´gico y de Estudios Superiores de Monterrey, Monterrey, Mexico Porfirio Caballero Center for Environmental Quality, Instituto Tecnolo´gico y de Estudios Superiores de Monterrey, Monterrey, Mexico Juan A. Villarreal and Ricardo Viramontes Ternium Hylsa, San Nicolas de los Garza, Nuevo Leon, Mexico

ABSTRACT A semi-industrial scale test was conducted to thermally treat mixtures of spent oil and askarels at a concentration of 50,000 ppm and 100,000 ppm of polychlorinated biphenyls (PCBs) under a reductive atmosphere. In average, the dry-basis composition of the synthesis gas (syngas) obtained from the gasification process was: hydrogen 46%, CO 34%, CO2 18%, and CH4 0.8%. PCBs, polychlorinated dibenzo-p-dioxins, and polychlorinated dibenzofurans (PCDDs/PCDFs) in the gas stream were analyzed by high-resolution gas chromatography (GC)-mass spectrometry. The coplanar PCBs congeners 77, 105, 118, 156/ 157, and 167 were detected in the syngas at concentrations ⬍2 ⫻ 10⫺7 mg/m3 (at 298 K, 1 atm, dry basis, 7% O2). The chlorine released in the destruction of the PCBs was transformed to hydrogen chloride and separated from the gas by an alkaline wet scrubber. The concentration of PCBs in the water leaving the scrubber was below the detection limit of 0.002 mg/L, whereas the destruction and removal efficiency was ⬎99.9999% for both tests conducted. The concentration of PCDDs/PCDFs in the syngas were 8.1 ⫻ 10⫺6 ng-toxic equivalent (TEQ)/m3 and 7.1 ⫻ 10⫺6 ng-TEQ/m3 (at 298 K, 1 atm, dry basis, 7% O2) for the tests at 50,000 ppm and 100,000 ppm PCBs, respectively. The only PCDD/F congener detected in the gas was the octachloro-dibenzo-p-dioxin, which has a toxic equivalent factor of 0.001. The results obtained for other pollutants (e.g., metals and particulate matter) meet the

IMPLICATIONS Gasification technology has in the past few years gained renewed interest because of its flexibility to be used in classical petrochemical processes, in the energy sector, or to valorize waste products. Here we present an application where gasification is used to treat contaminated oil with PCBs, obtaining high-destruction efficiencies and a commercially valuable effluent. The technology could be extended to treat other hazardous waste, making it a feasible alternative to incineration or other technologies.

Volume 56 November 2006

maximum allowed emission limits according to Mexican, U.S., and European regulations for the thermal treatment of hazardous waste (excluding CO, which is a major component of the syngas, and total hydrocarbons, which mainly represent the presence of CH4). INTRODUCTION Thermal destruction of hazardous and nonhazardous waste through incineration processes is a technology that has been used for some time. Incinerators are designed to optimize the possibility of complete oxidization of the waste and generate, mainly, CO2 and water in the combustion gases and ash in the remaining solid phase. An undesirable characteristic of incinerators is that they tend to produce and/or emit toxic compounds, such as heavy metals and polychlorinated dibenzo-p-dioxins (PCDDs) and polychlorinated dibenzofurans (PCDFs).1–5 Other thermal technologies are available to explore the treatment of hazardous and nonhazardous waste, like gasification.6 Thermal destruction of a carbonaceous material can occur under three main conditions based on the availability of oxygen: (1) combustion (oxidizing atmosphere), (2) gasification (limited oxygen producing a reductive atmosphere), and (3) pyrolysis (absence of oxygen). Gasification has multiple benefits that make it superior to incineration, the most important being the destruction of waste material and the production of a synthesis gas or syngas (CO ⫹ H2) that can be used as raw material for other processes or energy generation.6,7 An additional benefit is that the probability of producing PCDDs/PCDFs can be reduced because of the high temperatures and reductive atmosphere inside the equipment, particularly with waste having high-chlorine content. Pyrolysis has also being explored as an alternative to incineration.8 Depending on the feedstock, a pyrolysis process can produce three main products: char (or coke), pyrolysis oil, and synthesis gas, whereas gasification will only produce synthesis gas as its main product. Pyrolysis has been used, for example, for production of coke from plastics,9,10 syngas from easily degradable oil wastes,11 Journal of the Air & Waste Management Association 1599

Mendoza et al.

Figure 1. Simplified flow diagram of the gasification process: (1) auxiliary combustion chamber and (2) gasification burner.

and valuable liquid organic compounds from waste oils.8 When using pyrolysis to treat hard-to-degrade waste, typically a considerable amount of char is produced that can be of value or is sent to a gasification step to complete its destruction.12,13 Here, a high yield for synthesis gas was established and, thus, gasification was favored from pyrolysis. The treatment of waste contaminated with polychlorinated biphenyls (PCBs) is a serious problem around the world. Several technologies that depart from classic incineration have been published in the open literature to treat waste contaminated with organic compounds, and particularly PCBs, including steam plasma14 –16 and solvent extraction,17 followed by chemical dehalogenation or radiolitic dechlorination,18,19 chemical oxidation,20 –24 electrochemical dehalogenation,25 and biodegradation.26,27 In the United States, the U.S. Environmental Protection Agency (EPA) has considered an exclusion from the Resource Conservation and Recovery Act for secondary oilbearing refinery materials when processed in a gasification system.6,28 The industry has suggested that EPA include gasification of any carbonaceous material in such an exclusion. Here we explore the gasification of mixtures of oils contaminated with PCBs as an additional costeffective alternative to safely manage PCBs. The experiments were conducted in a facility located in Monterrey, Mexico. The facility used for the test is now a commercial site dedicated to treat spent oils contaminated with PCBs. Gasification Processes Gasification is a technology that thermally breaks down solid or liquid organic material (in fact, any carbonaceous material) to simple molecules (CO ⫹ H2) under an atmosphere poor in oxygen, that is, a reductive atmosphere, by supplying a gasification agent (typically steam).7,29 The gas stream produced is known as synthesis gas or syngas. The gasification process has been studied extensively, and, in practice, the syngas obtained is used as fuel to generate electricity and steam and as raw material in the production of a vast amount of chemical compounds, such as methanol and ammonia.6,30 If PCBs are present in the feedstock of a gasifier, a considerable amount of chlorine will be present in the 1600 Journal of the Air & Waste Management Association

reacting mixture, and, thus, the fate of such chlorine molecules is of interest because of their potential participation in the formation of PCDDs/PCDFs. It has been observed that gasification products in real applications tend to closely follow model predictions that assume thermodynamic equilibrium.31 If this is the case, one can argue that the reductive conditions in the gasification reactor thermodynamically favor the chlorine being present in the HCl form (i.e., the Deacon process reaction 2HCl[g] ⫹ 1⁄2O2[g] 7 Cl2[g] ⫹ H2O[g] will tend to the lefthand side product because of the absence of O2 and the presence of H2O as the gasifying agent), limiting the amount of Cl2 produced, and thus reducing the probability of chlorination of aromatic ring structures and consequently the formation of PCDDs/PCDFs.6,32 At the end of the process, gas-phase HCl present in the syngas can easily be absorbed in water to generate an aqueous solution of hydrochloric acid. In fact, the fate is the same for other organic compounds. Other authors have demonstrated or documented that gasification of chlorinated solvents33 and, in general, municipal solid waste34 and hazardous waste, is feasible,6 and in theory such compounds, like chlorofluorocarbons and pesticides, could also be treated by this technology. Process Description The gasification reactor used in this study is a 3.8-m3 vertical reactor, originally manufactured by M.W. Kellog, Co., and made of carbon steel ASTM 285 – Grade C, with four different insulation covers (two of concrete, refractory brick, and ceramic fiber). The total length of the reactor is 7.35 m. The main body of the reactor is a cylinder of 4.20 m in length and 0.89 m in i.d. On the extremes, the reactor narrows to accommodate on one side the burner (top) and on the other (bottom) a quencher that is incorporated to the exit pipeline. The design pressure is 1240 kPa, and the design temperature of the metal shell is 345 °C. The gasification burner is a device of proprietary design, specifically designed to insure proper mixing of the oxygen, steam (the gasifying agent), and oil (the carbon source) in the feed. Figure 1 depicts the gasification process. The system was started by enabling the cooling water circuit (closed Volume 56 November 2006

Mendoza et al. system) that provides water to different equipment, including the wet scrubber. Next, pressure tests were executed under nitrogen atmosphere to the reduction circuit followed by testing of the flow meters. The pressure in the gasification reactor (R-1) was maintained at 310 kPa throughout the operation. The start-up sequence continued with the preheating of the reactor, burning natural gas and oxygen in an auxiliary combustion chamber (located in 1 on Figure 1). At the same time, steam (the gasification agent) was fed to the gasification burner (located in 2 on Figure 1). When the temperature in R-1 reached the desired level, the gasification burner started to operate. To accomplish this, steam, oxygen, and spent oil were fed to the gasification burner. Once it was confirmed that the gasification burner was working, the gas burner was shut down. The temperature reading of the closest point in the wall to the burner was ⬃1140 °C at steady state. The residence time of the gases in the reactor was estimated in ⬃42 sec. The residence time was maintained high to give enough time to the feed to gasify and reduce the possibility of releasing undestroyed PCBs. At the exit of the reactor, the quencher reduced the temperature of the syngas, which was then sent to a scrubber before leaving the system. The temperature of the syngas leaving the quencher and the scrubber was ⬃40 °C and 35 °C, respectively. For the particular application described here, a significant amount of HCl was produced from the chlorine present in the PCB molecules and was absorbed in the water fed to the scrubber. If required, this HCl can be recovered from the solution for other applications. Here, the cooling/absorbing water had to be continuously neutralized with a sodium hydroxide solution because of the continuous accumulation of HCl in the water given the closed nature of the water circuit. The oil feeding system consisted of a series of tanks where the feed mixtures were prepared. The spent oil (without PCBs) was contained in a “day tank”; the oil received by this tank came from a main storage tank that collects oil from a variety of sources. The oil with PCBs was contained in a double-wall tank. The mixtures were prepared in a second double-wall tank. Oil flowed by gravity from the tanks containing the spent oil and the contaminated oil with PCBs to the mixing tank, and the required amounts were controlled by a series of valves connected to the distributed automatic control system of the plant. The amount required to be mixed was controlled by measuring the added weight from each tank. The feed mixture was then sent by gravity to the feeding tank, where finally the mixture was pumped to R-1. EXPERIMENTAL WORK The plant started operations in the morning of May 11, 2003, and reported stable conditions in the reactor in the afternoon of that same day. Sampling of the exit gas stream, the feedstock, and the water streams started on May 12 and continued until May 14. The feed of the first day was spent oil with no PCBs, the second day the oil was prepared to have 5% PCBs, and the third day the feed had 10% PCBs. The spent oil was a mixture of residual hydraulic oils from typical process equipments (pumps, compressors, mills, etc.), whereas the oil contaminated with PCBs came from unused electric transformers. To minimize Volume 56 November 2006

risks, the feed of PCBs was suspended during the break between the sampling of the gas stream under the 5% and 10% PCBs conditions. After completing the test with oil containing 10% PCBs, uncontaminated oil was once again fed to the system, which remained operating 24 additional hours to eliminate any residues of PCBs in the equipment. The syngas produced was sampled and its composition analyzed, as well as the oil mixtures fed to the reactor and the quality of the water leaving the scrubber. Figure 1 illustrates the location of the water and gas streams sampled. The gas was analyzed for CO2, CO, O2, N2, H2, CH4, nitrogen oxide (NOx), SO2, HCl, total hydrocarbons (THCs), total suspended particulate (TSP) matter, metals (As, Cd, Co, Cr, Cu, Hg, Mn, Ni, Pb, Se, Sn, and Zn), PCBs, and PCDDs/PCDFs. The feed was analyzed for PCBs, and the water was analyzed for PCBs, chlorides, and PCDDs/ PCDFs. The samples were obtained and analyzed following standard Mexican methods and, where applicable, EPA methods (Table 1). In particular, for PCBs and PCDDs/ PCDFs in the stack, EPA Method 23 was used, implemented with an isokinetic source sampling system (Environmental Supply Company, Inc., model CS-5000) as described in EPA Method 5 and augmented with Method 23 extension glassware. The preparation of samples and posterior analysis was conducted by Alta Analytical Perspectives. Capture resins (XAD-2) were spiked with surrogate standards as follows: for PCDD/PCDF, 20 ␮L of a 0.2-ng/␮L solution was used, whereas for PCBs, 40 ␮L of a 0.1-ng/␮L solution was used. Capture solutions contained high-performance liquid chromatography-grade solutions of water, methylene chloride, toluene, and acetone. A high-resolution gas chromatograph/high-resolution mass spectrometer (VG 70SE High-Resolution Double Focusing Mass Spectrometer) system was used to analyze the extracts obtained. Method 1668A, a water method modified by the analytical laboratory, was used in the highresolution analysis of PCBs in the stack. Three 1-hr samples and a field blank were obtained for each operation condition, and an average of 1.2 m3 passed through the sampling train. The reported values were calculated as an average of the three samples, corrected by the value obtained for the field blank. Sampling for each operation condition started once the system was considered to have stabilized (⬃1 hr after the transition to the operation condition). Metals/TSP, HCl, and SO2 samples were also collected using isokinetic source sampling systems. As for the previous case, for metals/TSP, three 1-hr samples and a field blank were obtained for each operation condition, and for HCl and SO2, two 1-hr samples and a field blank were collected. An average of 1.2 m3 was also sampled for these species. Metals were analyzed by Inductively Coupled Plasma Atomic Emission Spectrometry (Thermo Jarrell Ash Corp., model Atom Scan 16), except Hg, which was analyzed by Cold Vapor Atomic Absorption (Varian Spectra, model Spectra AA). HCl and SO2 were analyzed by ion chromatography ([IC] Dionex, model DX-100). NOx, O2, CO, CO2, and THC were measured semicontinuously during the whole sampling period. NOx was Journal of the Air & Waste Management Association 1601

Mendoza et al. Table 1. Sampling and analysis methods used. Source Stack gas

Liquid samples

Method NMX-AA-09-1993-SCFI NMX-AA-54-1978 NMX-AA-10-SCFI-2001, EPA 5 NMX-AA-55-1979 EPA 3A, EPA 10 EPA 25A EPA 26A EPA 7E EPA 23 EPA 29 EPA 300.1-1999 EPA 8081-1994 EPA 9253

monitored using a chemiluminescence device (ThermoEnvironmental Instruments, Inc., model 42-H), O2 by electrochemical cells (Bacharach, model PCA), CO and CO2 by nondispersive infrared technique (Milton Roy, model ZRH), and THC by flame ionization (Rosemount Analytical, model 400A). In addition, instantaneous samples were collected on an hourly basis in metal bulbs and analyzed for H2, N2, CO, CO2, and CH4 by GC (conductivity detector). The results for CO and CO2 compared well with the continuous samples. Finally, PCBs in water samples and feed oil were analyzed by GC-electron-capture detector (Agilent, model HP5890), chlorides in water by IC, and PDCC/PDCFs by high-resolution GC-mass spectrometry as the stack samples. Feed oil samples were a composite obtained through the corresponding operation condition period, whereas water samples were instantaneous samples collected at the middle of each condition period. RESULTS Ultimate analysis of the dilution oil indicated an average composition of 83.4% carbon, 10.5% hydrogen, 3.1% ash, 1.8% oxygen, 0.6% sulfur, and 0.6% humidity. Trace quantities of chlorine (0.05%) were also present. The heating value obtained for the oil was 10,710 kcal/kg. Chemical analysis of the transformer oil contaminated with PCBs indicated a heating value of 5150 kcal/kg, a content of PCBs of 66.8% and 40% of chlorine. The characterizations of the mixtures fed to R-1 are reported in Table 2. The heating value of the mixture decreased, as expected, as more PCBs were present. Regarding the metals, Zn was the most abundant of those analyzed, whereas As, Cd, Co, Se, Sn, and Hg were not detected. In all of the cases, the content of metals decreased as the PCBs content increased because of dilution of the spent oil with the contaminated oil with PCBs, which had less metal content. Of note, given that the experiments were conducted at a semi-industrial scale, during different days, with spent oil acquired from different sources and, thus, with different quality related to the amount of trace metals present, it is not expected that the amount of trace metals correlates well with the dilution factors of each test (i.e., 5% and 10% dilution). The appropriate amount of spent oil and contaminated oil used 1602 Journal of the Air & Waste Management Association

Description Determination Determination Determination Determination Determination Determination Determination Determination Determination Determination Determination Determination Determination

of of of of of of of of of of of of of

stack gas velocity and volumetric flow rate (pitot tube method) stack gas humidity content (gravimetric method) particulate emissions from stationary sources sulfur dioxide CO2, CO, and O2 (instrumental) total hydrocarbons (flame ionization) halogens and hydrogen halides NOx (chemiluminescence) PCDD/PCDFs metals inorganic anions (chlorides) in water (IC) organochloride pesticides and PCBs as Aroclor (GC–capilar column) chlorine in aqueous solution

to generate each mixture was corroborated with the average amount of PCBs actually present in the feed to the reactor. The concentration of PCBs in the feed that contained only spent oil was ⬍0.4 mg/kg; for the 5% condition the actual average concentration was 4.9%, and for the 10% condition it was 11%. Table 3 summarizes the amount of oil (with and without PCBs), steam, and oxygen fed during the gasification tests, whereas Table 4 presents the average composition of the syngas obtained. As shown, the syngas obtained is of very good quality with a high concentration of reducers (H2 and CO). In general, the syngas obtained has a higher concentration of H2 and less CO2 compared with the one produced by other authors for the gasification of municipal solid waste35 or motor oil,31 mainly because of differences in operating temperature and chemical composition of the feed. The result obtained for HCl is of particular interest, because the analyses indicate concentrations lower than the detection limit of the method (⬍0.03 mg/m3). As indicated earlier, thermodynamic considerations31 indicate that the chlorine present in the PCBs will preferably be transformed into HCl. This Table 2. Average chemical characterization of the mixtures fed to the gasification reactor by operation condition. Parameter

Units

0% PCBs

5% PCBs

10% PCBs

Heating value Chlorine Sulfur Water As Cd Co Cr Cu Mn Ni Pb Se Sn Zn Hg

kcal/kg % % % mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg

10,710 0.05 0.61 0.60 ⬍7 ⬍0.4 ⬍1 10 23 50 4 14 ⬍8 ⬍5 1370 ⬍0.05

10,280 3.23 0.64 0.54 ⬍7 ⬍0.4 ⬍1 ⬍5 18 44 ⬍2 13 ⬍8 ⬍5 1270 ⬍0.05

9850 5.28 0.53 0.53 ⬍7 ⬍0.4 ⬍1 ⬍5 16 41 ⬍2 12 ⬍8 ⬍5 1220 ⬍0.05

Volume 56 November 2006

Mendoza et al. Table 3. Summary of average oil, steam, and oxygen feeding rates to the gasification system. Oil Consumption (kg/hr)

Operation Condition

Operation Time of the Operation Condition (hr)

Steam (scmh)

Oxygen (scmh)

Dilution Oil (0% PCBs)

13.0 24.0 14.0 10.0 14.0 24.0

72.6 74.8 73.4 71.8 77.0 77.8

61.9 62.8 61.5 62.7 62.8 63.0

56.0 56.9 53.9 57.0 51.7 56.7

Start-up 0% PCBs in the oil feed 5% PCBs in the oil feed Intermediate condition (0% PCBs in the oil feed)a 10% PCBs in the feed Final condition/cleanup (0% PCBs in the oil feed)a

Oil with PCBs (66.83% PCBs)

4.3 8.9

Notes: scmh ⫽ standard cubic meters per hour. aNo sampling conducted during this operation condition.

is in part corroborated by experimental results from the gasification of other types of waste.32 That being the case, the fate of the HCl would be in the water of the scrubber, as was the case. Analysis of particulate matter and heavy metals in the clean syngas also indicated low concentrations for these species. Total particulate matter concentrations were 6.3, 2.1, and 2.7 mg/m3 for the conditions of 0%, 5%, and 10% PCBs, respectively. Pb, Cr, Cu, As, Se, Co, and Cd were not detected on the syngas. The rest of the metals reported their highest concentration under the 0% PCB condition (Zn 0.127, Ni 0.002, Mn 0.002, and Sn 0.078 mg/m3), with the exception of Hg, which had its highest concentration (0.0002 mg/m3) under the 10% PCB condition. These results are in general agreement with the metal content in the oil fed in each test. Results obtained here corroborate experimental and modeling results that indicate that a wet scrubbing system should be effective for removing these elements from the gas (except Hg, which in any case was undetected in the feed).36 All of the values reported for particulate matter and metals are at normal conditions of temperature and pressure (298 K, 1 atm), dry basis, and 7% O2. Detailed results of PCBs in the syngas are reported in Figure 2. Highest concentrations were reported for triPCBs, although in the test with 10% PCBs in the feed, mono-PCBs and di-PCBs dominated the total concentration. Speciation of PCBs by congeners is reported in Table Table 4. Characterization of the syngas for each operation condition. Parameter

Unitsa,b

0% PCBs

5% PCBs

10% PCBs

H2 CO CO2 O2 CH4 N2 NOx SO2 HCl Total PCBs PCDDs/Fs

%V %V %V %V %V %V mg/m3 mg/m3 mg/m3 mg/m3 ng-TEQ/m3

46.7 33.4 18.7 0.05 1.0 0.2 1.7 1.9 ⬍0.03 2.2 ⫻ 10⫺5 6.5 ⫻ 10⫺6

46.1 34.2 18.5 0.09 0.9 0.3 1.9 2.2 ⬍0.03 3.4 ⫻ 10⫺5 8.1 ⫻ 10⫺6

45.6 34.6 18.9 0.09 0.6 0.4 2.8 1.1 ⬍0.03 9.0 ⫻ 10⫺4 7.1 ⫻ 10⫺6

Notes: aPercent volume is on a dry basis; bmg or ng per m3 are at 298 K, 1 atm, dry basis, and corrected to 7% O2. Volume 56 November 2006

5. The selected congeners are considered of equivalent toxicity to PCDDs/PCDFs because of their coplanar molecular structure. The detected congeners have toxicity equivalent factors (TEFs) in the range of 0.0001 to 0.1 with respect to the most toxic known dioxin (2,3,7,8 tetrachloro dibenzo-p-dioxins). Of the listed species, only PCB-77, -105, -118, -156/157, and -167 were detected, and PCB-118 (TEF ⫽ 0.0001) was the one reported with the highest concentration in all of the operation conditions. A similar analysis was conducted for PCDDs/PCDFs congeners, although the only species detected was the octachloro-dibenzo-p-dioxin (OCDD), which has a TEF of 0.001. OCDD concentrations and corresponding relative standard deviations were 6.5 ⫻ 10⫺6 (64%), 8.1 ⫻ 10⫺6 (120%), and 7.1 ⫻ 10⫺6 ng/m3 (90%) TEQ (normal conditions, dry basis, 7% O2), under the 0%, 5%, and 10% PCBs conditions, respectively. Mean recovery of extraction standards for the OCDD was 84% (N ⫽ 13), and in general the mean recoveries were in the range of 81–90%. The water used in the scrubber/quencher was analyzed for PCBs, dissolved chlorine (Cl2), chlorides, and PCDDs/PCDFs, and the results are shown in Table 6. Dissolved Cl2 and PCBs were below the detection limit of the method in all of the cases (0.09 and 0.002 mg/L, respectively). The increment in the concentration of chlorides is because of the closed nature of the cooling water system, thus, accumulation of chlorides and PCDDs/PCDFs is observed. As discussed earlier, HCl is produced in the reactor, which is then captured in the scrubber. The only PCDDs/PCDFs detected were the OCDD (as in the case of the syngas samples) and 2,3,7,8-tetra-chloro-dibenzofuran, with TEF of 0.001 and 0.1, respectively. The rest of the congeners were not detected. The European Cooperation Community rules (CE 2000/C25-02) consider a reference limit of 0.3 ng-TEQ/L of PCDDs/Fs in the water from control devices of hazardous treatment processes. The maximum concentration obtained in the test was 0.0084 ng-TEQ/L. DISCUSSION A feasibility study on the gasification of PCBs was presented. The results obtained indicate that the technology has the potential to be used in the conversion of contaminated oils with askarels to a clean syngas. The test demonstrated that small quantities of pollutants, and particularly air toxics, are present in the syngas. Polyaromatic hydrocarbon formation, Journal of the Air & Waste Management Association 1603

Mendoza et al.

Figure 2. PCBs characterization in the syngas by PCB class for each operation condition (concentrations are at normal conditions, dry basis, and corrected to 7% O2).

which could be of concern in the ash as demonstrated in the gasification of other types of waste,37 was not addressed. During the tests conducted, the oil feed rate increased from 56.9 to 60.6 kg/hr, which, in turn, represented a decrease in the heating value and an increase in the chlorine of the feed mixture, with a corresponding decrease in the relative amounts of carbon and hydrogen present in the oil. Overall, the mixture became harder to gasify as more PCBs were present. This condition had to be compensated with an increase in the steam-oxygen volumetric flow ratio, which started at 1.19 and ended at 1.23, because steam increases the reactivity of the gasifying mixture.31,38 With this, more gasifying agent was introduced to insure proper destruction of the feedstock, gasification of carbon, and limit to the production of particulate matter. A consequence of having more PCBs in

the feedstock was the decrease of H2 and CH4 being produced and an increment of CO and CO2. Also, the nondestroyed amounts of PCBs in the syngas increased, particularly under the condition with a feed of 10% PCBs. The profile of PCBs by class (Figure 2) under the 5% condition resembles well background (blank) conditions; however, mono-PBCs and di-PBCs are present with concentrations of ⬎2 orders of magnitude larger under the 10% condition. Although the total amount of ungasified PCBs increased by ⬎1 order of magnitude during the last operation condition, the destruction efficiency was still very high, as discussed later. Finally, decreasing PCDD/ PCDF concentrations correlated well (R2 ⫽ 0.77) with increasing steam-to-oil feed ratios indicating a decreasing tendency to produce these toxic compounds under more reductive conditions in the gasification reactor.

Table 5. PCB congeners in the syngas (all concentrations are reported as mg/m3 at 298 K, 1 atm, dry basis, and 7% O2).

Congener PCB-77 PCB-81 PCB-105 PCB-114 PCB-118 PCB-123 PCB-126 PCB-156/157 PCB-167 PCB-169 PCB-189

Toxicity Equivalent Factors (WHO mammals/humans)

0% PCBs (mg/m3)

5% PCBs (mg/m3)

10% PCBs (mg/m3)

0.0001 0.0001 0.0001 0.0005 0.0001 0.0001 0.1 0.0005 0.00001 0.01 0.0001

4.7 ⫻ 10⫺8 ⬍3.5 ⫻ 10⫺9 4.8 ⫻ 10⫺8 ⬍3.0 ⫻ 10⫺9 1.8 ⫻ 10⫺7 ⬍3.2 ⫻ 10⫺9 ⬍1.7 ⫻ 10⫺9 6.2 ⫻ 10⫺9 1.4 ⫻ 10⫺9 ⬍2.5 ⫻ 10⫺9 ⬍3.3 ⫻ 10⫺9

3.0 ⫻ 10⫺8 ⬍3.4 ⫻ 10⫺9 2.5 ⫻ 10⫺8 ⬍3.1 ⫻ 10⫺9 8.5 ⫻ 10⫺8 ⬍3.2 ⫻ 10⫺9 ⬍2.8 ⫻ 10⫺9 9.6 ⫻ 10⫺10 ⬍1.9 ⫻ 10⫺9 ⬍2.7 ⫻ 10⫺9 ⬍3.4 ⫻ 10⫺9

4.2 ⫻ 10⫺8 ⬍3.0 ⫻ 10⫺9 3.5 ⫻ 10⫺8 ⬍3.4 ⫻ 10⫺9 9.6 ⫻ 10⫺8 ⬍3.6 ⫻ 10⫺9 ⬍4.5 ⫻ 10⫺9 5.7 ⫻ 10⫺9 1.7 ⫻ 10⫺9 ⬍2.2 ⫻ 10⫺9 ⬍3.1 ⫻ 10⫺9

Notes: WHO ⫽ World Health Organization. 1604 Journal of the Air & Waste Management Association

Volume 56 November 2006

Mendoza et al. Table 6. Characterization of the process water.

Condition

Chlorides (mg/L)

Dissolved Cl2 (mg/L)

PCBs (mg/L)

PCDDs/Fs (ng-TEQ/L)

Start-up 0% PCBs 5% PCBs 10% PCBs Cleaning

59 74 756 1850 2551

⬍0.09 ⬍0.09 ⬍0.09 ⬍0.09 ⬍0.09

⬍0.002 ⬍0.002 ⬍0.002 ⬍0.002

9.4 ⫻ 10⫺5 1.5 ⫻ 10⫺4 8.4 ⫻ 10⫺3

The performance of the gasification process was compared with emission rules available in Mexico, the United States, and Europe. It needs to be noted that the rules were intended for application to the combustion gases emitted by incineration processes. In the case of a gasification process, the synthesis gas obtained is not emitted to the atmosphere, because it can be further used in other processes. Thus, it is no longer a residue. The comparison presented here used the rules for incineration only as a guideline. Table 7 presents the comparison of the syngas composition with respect to Mexico’s MRP-6 official document on hazardous waste treatment processes. CO data were not included, because it is precisely the objective of the gasification process to produce this compound. In the same line, most of the THC is present as CH4. These two compounds are not residues of the process, but products that can be used in other processes. Table 7 indicates compliance of the rest of the parameters with respect to the Mexican rule. Emissions of PCDDs/PCDFs indicate 5 orders of magnitude lower than the maximum emission limit. The measured values for PCDDs/PCDFs of ⬃0.01 ng/m3 (normal and dry conditions) were further reduced because of the small amount of O2 present in the exit stream (a factor of 0.67 because of the correction to report at 7% O2) and because of the fact that the only PCDD/ PCDF detected was the OCDD, which has a toxic equivalent correction factor of 0.001. Comparisons of the levels of PCDDs/PCDFs obtained in this work with values reported in the literature indicate lower concentrations for these tests. For example, lower-limit levels of 0.001 ng/m3 TEQ have been reported in the gasification of highly

chlorinated feedstocks,6 and the presence of OCDD in flue gas from the gasification of municipal solid waste has been reported at 0.17 ng/m3 (at 11% O2).34 The destruction and removal efficiency, with respect to the total amount of PCBs fed, was ⱖ1 order of magnitude above the minimum allowable. A similar analysis was conducted using as guidelines Title 40 of the U.S. Code of Federal Regulations, Part 60 and Common Position (CE) No. 7/2002, Annex V of the European Community. In 1999, EPA published the document “NESHAPS: Final Standards for Hazardous Air Pollutants for Hazardous Waste Combustors, Final Rule,” based on evaluations of the maximum Achievable Control Technology Rule. The maximum allowable emission limits were established at standard conditions (293 K, 1 atm), dry basis, and 7% O2. The difference in temperature with respect to the Mexican rule produces a correction factor of 1.017 in the concentration reported in Table 7. The maximum allowable emission limits by the European regulation cited above were established using the following conditions: dry basis, 273 K, 101.3 kPa, and 11% oxygen. The temperature produces a correction factor of 1.091; meanwhile, the oxygen produces a correction factor of 0.784 from the values reported in Table 7. Thus, a comparison of measured values against the limits imposed by U.S. and European regulations can be obtained directly and is not shown for brevity. The results obtained from this analysis indicate that the syngas obtained would meet the maximum allowed emission limits set by Mexican, U.S., and European regulations for the thermal treatment of hazardous waste for the following parameters: particulate matter, NOx, SO2, HCl, metals, PCCDs/PCDFs, and destruction efficiency. Because the syngas would not be emitted, but instead used by other processes, the results indicate that this raw material is of high quality and free of those hazardous pollutants analyzed. The levels of NOx and SO2 are also very low thanks to the reductive atmosphere present in the reactor that, instead of oxidizing the nitrogen and sulfur present in the fuel, it reduces them to N2 and H2S. A benefit from this is that for fuels with high sulfur content, the H2S can be easily captured and transformed to elemental sulfur or can produce sulfuric acid solution.

Table 7. Comparison of results of the test with rules available in Mexico. Parameter Total suspended particles CO NOx SO2 HCl THC Pb ⫹ Cr ⫹ Cu ⫹ Zn As ⫹ Se ⫹ Co ⫹ Ni ⫹ Mn ⫹ Sn Cd Hg PCDD/Fs Destruction and removal efficiency

Unitsa mg/m3 mg/m3 mg/m3 mg/m3 mg/m3 mg/m3 mg/m3 mg/m3 mg/m3 mg/m3 ng-TEQ/m3 %

0% PCBs

5% PCBs

10% PCBs

6.3

2.1

2.7

1.7 1.9 ⬍0.03

1.9 2.2 ⬍0.03

2.8 1.1 ⬍0.03

0.127 0.082 ⬍0.0002 0.00004 6.5 ⫻ 10⫺6

0.067 0.071 ⬍0.0002 0.00005 8.1 ⫻ 10⫺6 ⬎99.9999

0.041 0.047 ⬍0.0002 0.00019 7.1 ⫻ 10⫺6 ⬎99.9999

Limit 30 63 300 80 15 10 0.7 0.7 0.07 0.07 0.5 99.9999 (minimum)

Notes: aConcentrations are reported at 298 K, 1 atm, dry basis, and 7% O2. Volume 56 November 2006

Journal of the Air & Waste Management Association 1605

Mendoza et al. CONCLUSIONS Results from a test on the gasification of spent oil contaminated with PCBs were presented. Tests were conducted using oil mixtures containing 0%, 5%, and 10% PCBs. The synthesis gas produced by the reactor was analyzed for gases, particles, metals, PCBs, and PCDDs/ PCDFs. The water used in the scrubber was analyzed for PCBs, chlorine, chlorides, and PCDDs/PCDFs. The technology efficiently destroyed PCBs contained in the feed (efficiency ⬎99.9999%), leaving most of the chlorine in the HCl form, which was captured in solution in the wet scrubber. The levels of the species sampled in the exit syngas were below the emission limits stated by the applicable Mexican, U.S., and European rules, including PCDDs/PCDFs (excluding CO and THC, the first a major product expected and the second a byproduct that is mostly CH4). The resulting syngas was of high quality, with high proportions of CO and H2 and free of those hazardous pollutants that were analyzed. This product can be further used in other synthesis processes or in energy generation equipment. REFERENCES 1. Eklund, G.; Pedersen, J.; Stro ¨ mberg, B. Phenol and HCl at 550 °C Yield a Large Variety of Chlorinated Toxic Compounds; Nature 1986, 320, 155-156. 2. Cudahy, J.J.; Rigo, H.G. National Annual Dioxin Emissions Estimate for Hazardous Waste Incinerators; J. Air & Waste Manage. Assoc. 1998, 48, 1107-1111. 3. Olie, K.; Addink, R.; Schoonenboom, M. Metals as Catalysts during the Formation and Decomposition of Chlorinated Dioxins and Furans in Incineration Processes; J. Air & Waste Manage. Assoc. 1998, 48, 101105. 4. Everaert, K.; Baeyens, J. Correlation and PCDD/F Emissions with Operating Parameters of Municipal Solid Waste Incinerators; J. Air & Waste Manage. Assoc. 2001, 51, 718-724. 5. Oh, J.-E.; Chang, Y.-S.; Ikonomou, M.G. Levels and Characteristic Homologue Patterns of Polychlorinated Dibenzo-p-Dioxins and Dibenzofurans in Various Incinerator Emissions and Air Collected Near an Incinerator; J. Air & Waste Manage. Assoc. 2002, 52, 69-75. 6. Orr, D.; Maxwell, D.A. Comparison of Gasification and Incineration of Hazardous Wastes; DCN 99.803931.02; Final Report Prepared for National Energy Technology Laboratory, U.S. Department of Energy: Morgantown, WV; by Radian International LLC: Austin, TX, 2000. 7. Belgiorno, V.; De Feo, G.; Della Rocca, C.; Napoli, R.M.A. Energy from Gasification of Solid Wastes; Waste Manage. 2003, 23, 1-15. 8. Moliner, R.; La´zaro, M.; Suelves, I. Valorization of Lube Oil Waste by Pyrolysis; Energy Fuels 1997, 11, 1165-1170. 9. Jiu-ju, C.; Guang-wei, Y.; Hong-qiang, L.; Kai, Q.; Pen, Z.; Ya-bin, H. Disposal of Waste Plastics with Traditional Coking Process; J. Iron Steel Res. Int. 2006, 13, 5-9. 10. Cozzani, V. Characterization of Coke Formed in the Pyrolysis of Polyethylene; Ind. Eng. Chem. Res. 1997, 36, 5090-5095. 11. Yang, H.; Yan, R.; Chin, T.; Liang, D.T.; Chen, H.; Zheng, C. Thermogravimetric Analysis-Fourier Transform Infrared Analysis of Palm Oil Waste Pyrolysis; Energy Fuels 2004, 18, 1814-1821. 12. Haykiri-Acma, H.; Yaman, S.; Kucukbayrak, S. Gasification of Biomass Chars in Steam-Nitrogen Mixture; Energy Conversion Manage. 2006, 47, 1004-1013. 13. Sainz-Diaz, C.I.; Kelly, D.R.; Avenell, C.S.; Griffiths, A.G. Pyrolysis of Furniture and Tire Wastes in a Flaming Pyrolyzer Minimizes Discharges to the Environment; Energy Fuels 1997, 11, 1061-1072. 14. Di, P.; Chang, D.P.Y. Investigation of Polychlorinated Biphenyl Removal from Contaminated Soil Using Microwave-Generated Steam; J. Air & Waste Manage. Assoc. 2001, 51, 482-488. 15. Kim, S.-W.; Park, H.-S.; Kim, H.-J. 100 kW Steam Plasma Process for Treatment of PCBs (Polychlorinated Biphenyls) Waste; Vacumm 2003, 70, 59-66. 16. Ko, Y.; Yang, G.; Chang, D.P.Y., Kennedy, I.M. Microwave Plasma Conversion of Volatile Organic Compounds; J. Air & Waste Manage. Assoc. 2003, 53, 580-585. 17. Meckes, M.C.; Tillman, J.; Drees, L.; Saylor, E. Removal of PCBs from a Contaminated Soil Using cf-Systems® Solvent Extraction Process; J. Air & Waste Manage. Assoc. 1997, 47, 1119-1124. 18. Nam, P.; Kapila, S.; Liu, Q.; Tumiatti, W.; Porciani, A.; Flanigan, V. Solvent Extraction and Tandem Dechlorination for Decontamination of Soil; Chemosphere. 2001, 43, 485-491. 1606 Journal of the Air & Waste Management Association

19. Chaychian, M.; Jones, C.; Poster, D.; Silverman, J.; Neta, P.; Huie, R.; Al-Sheikhly, M. Radiolytic Dechlorination of Polychlorinated Biphenyls in Transformer Oil and Marine Sediment; Radiat. Phys. Chem. 2002, 65, 473-478. 20. Ruggeri, B.; Tundo, P.; Tumiatti, W. Supported Liquid Phase Reactor (SLPR) for PCBs in Oil Decontamination; Chem. Eng. Sci. 1990, 45, 2867-2693. 21. Ryo, K.S.; Kapila, S.; Puri, R.K.; Yanders, A.F.; Elseewi, A.A. Evaluation of Carbon for Removal and Destruction of Polychlorinated Biphenyls (PCBs) from Transformer Mineral Oils; Chemosphere 1992, 25, 15691575. 22. De Fillipis, P.; Chianese, A.; Pochetti, F. Removal of PCBs from Mineral Oils; Chemosphere 1997, 35, 1659-1667. 23. Lindsey, M.E.; Xu, G.; Lu, J.; Tarr, M.A. Enhanced Fenton Degradation of Hydrophobic Organics by Simultaneous Iron and Pollutant Complexation with Cyclodextrins; Sci. Total Environ. 2003, 307, 215-229. 24. Seok, J.; Seok, J.; Hwang, K. Thermo-Chemical Destruction of Polychlorinated Biphenyls (PCBs) in Waste Insulating Oil; J. Haz. Mat. 2005, 124, 133-138. 25. Rundhaug, P. On-Site Electrochemical Dehalogenation Process and System; J. Cleaner Production 1996, 4, 252. 26. Rojas-Avelizapa, N.G.; Rodrı´guez-Va´zquez, R.; Enrı´quez-Villanueva, F.; Martı´nez-Cruz, J.; Poggi-Varaldo, H.M. Transformer Oil Degradation by an Indigenous Microflora Isolated from a Contaminated Soil; Resour. Conserv. Recycl. 1999, 27, 15-26. 27. Sierra, I.; Valera, J.L.; Marina, M.L.; Laborda, F. Study of the Biodegradation Process of Polychlorinated Biphenyls in Liquid Medium and Soil by a New Isolated Aerobic Becterium (Janibacter sp.); Chemosphere 2003, 53, 609-618. 28. Shirley, W.A. Regulatory Update: EPA to Allow Petroleum Wastes as an Energy Source; Chem. Eng. Prog. 2002, 98, 22. 29. Di Blasi, C. Dynamic Behaviour of Stratified Downdraft Gasifier; Chem. Eng. Sci. 2000, 55, 2931-2944. 30. Considine, D.M. Energy Technology Handbook; McGraw-Hill: New York, NY, 1977. 31. Larsen, D.W.; Washington, M.D.; Manahan, S.; Medcalf, B.; Stary, F. Thermodynamic Considerations in the Application of Reverse Mode Gasification to the Destruction of Hazardous Substances; Environ. Sci. Technol. 1999, 33, 2973-2979. 32. Bjo ¨ rkman, T.; Stro ¨ mberg, B. Release of Chlorine from Biomass at Pyrolysis and Gasification Conditions; Energy Fuels 1997, 11, 10261032. 33. Morlando, R.; Manahan, S.E.; Larson, D.W. Iron-Catalyzed Cocurrent Flow Destruction and Dechlorination of Chlorobenzene during Gasification; Environ. Sci. Technol. 1997, 31, 409-415. 34. Liu, Y.; Liu, Y. Novel Incineration Technology Integrated with Drying, Pyrolysis, Gasification, and Combustion of MSW and Ashes Vitrification; Environ. Sci. Technol. 2005, 39, 3855-3863. 35. Kwak, T.-H.; Lee, S.; Maken, S.; Shin, H.-C.; Park, J.-W.; Yoo, Y.D. A Study of Gasification of Municipal Solid Waste Using a Double Inverse Diffusion Flame Burner; Energy Fuels 2005 19, 2268-2272. 36. Reed, G.P.; Paterson, N.P.; Zhuo, Y.; Dugwell, D.R.; Kandiyoti, R. Trace Element Distribution in Sewage Sludge Gasification: Source and Temperature Effects; Energy Fuels 2005, 19, 298-304. 37. Asikainen, A.J.; Kuusisto, M.P.; Hiltunen, M.A.; Ruuskanen, J. Ocurrence and Destruction of PAHs, PCBs, CIPhs, CIBzs, and PCDD/Fs in Ash from Gasification of Straw; Environ. Sci. Technol. 2002, 36, 21932197. 38. Mu ¨ hlen, H.J.; van Heek, K.H.; Ju ¨ ntgen, H. Kinetic Studies of Steam Gasification of Char in the Presence of H2, CO2 and CO; Fuel 1985, 64, 944-949.

About the Authors Alberto Mendoza is an associate professor in the Department of Chemical Engineering, Instituto Tecnologico y de Estudios Superiores de Monterrey (ITESM). Porfirio Caballero is an associate professor in the Center for Environmental Quality at ITESM. Juan A. Villarral and Ricardo Viramontes are with Ternium Hylsa. Address correspondence to: Alberto Mendoza, Department of Chemical Engineering, Instituto Tecnologico y de Estudios Superiores de Monterrey (ITESM), Ave. Eugenio Garza Sada 2501 Sur, Monterrey, Nuevo Leon 64849, Mexico; phone: ⫹52-818328-4336; fax: ⫹52-81-8328; e-mail: mendoza.alberto@ itesm.mx.

Volume 56 November 2006