《纺织复合材料》课程参考文献（Fibers and Composites）07 LIQUID IMPREGNATION TECHNIQUES FOR CARBON–CARBON COMPOSITES.pdf_大学文库

7 LIQUID IMPREGNATION TECHNIQUES FOR CARBON-CARBON COMPOSITES R.Menendez,E.Casal,and M.Granda 1 Introduction Carbon fiber reinforced carbon composites (C-C composites)are unique in providing materials with the highest possible specific thermal-mechanical properties.They have an additional value in that at high temperatures their excellent properties are not only retained but also improved as a result of a better structural order.However,the main problem with C-C composites,apart from their reactivity to oxygen at high temperatures,is their high cost due to the complexity and length of the fabrication processes.Consequently,C-C composites are currently used only in high-technology applications where no other materi- als are available.In recent years the search for ways to improve properties,reduce costs,and increase the efficiency of the processes involved in the preparation of C-C composites has been intensified.Unfortunately,easily available information is very limited because of con- fidentiality agreements,most of it being in patent form(Plotzker et al.,1988;Hocquellet and Chenier,1993). The type and architecture of the fibers and the microstructure of the supporting matrix determine the physical properties of C-C composites.It is,therefore,important to under- stand the mechanisms that control the formation of the microstructure and its interaction with the fibers in order to be able to achieve the desired composite properties at the lowest possible cost and make full use of the electrical,thermal,and mechanical properties of the fibers (Meyer,1994). Two methods are at present used for the preparation of C-C composites:(i)liquid impreg- nation;and(ii)carbon vapor infiltration (CVI).A combination of the two methods is also used.Both have been described in detail in the literature(Savage,1993a).Liquid impreg- nation,the topic covered in this chapter,is a two-stage method.Basically,in the first stage the fibers are impregnated with an organic precursor,giving rise to what is called a prepreg. Afterwards,the precursor is transformed into a carbonaceous matrix by thermal treatment (650-1,000C)in an inert atmosphere via a carbonization process.Depending on the appli- cation of the material,graphitization (>2,500C)may also be required.These stages have to be repeated until the desired density is obtained (Fig.7.1).The other method,CVI,implies the cracking of a gas and the deposition of carbon onto the fiber substrate (Chapter 7 of this book).In general,liquid impregnation is considerably quicker and cheaper than CVI pro- cessing,and the equipment required,in an ambient pressure fabrication route,is not limited ©2003 Taylor&Francis

7 LIQUID IMPREGNATION TECHNIQUES FOR CARBON–CARBON COMPOSITES R. Menéndez, E. Casal, and M. Granda 1 Introduction Carbon fiber reinforced carbon composites (C–C composites) are unique in providing materials with the highest possible specific thermal-mechanical properties. They have an additional value in that at high temperatures their excellent properties are not only retained but also improved as a result of a better structural order. However, the main problem with C–C composites, apart from their reactivity to oxygen at high temperatures, is their high cost due to the complexity and length of the fabrication processes. Consequently, C–C composites are currently used only in high-technology applications where no other materials are available. In recent years the search for ways to improve properties, reduce costs, and increase the efficiency of the processes involved in the preparation of C–C composites has been intensified. Unfortunately, easily available information is very limited because of confidentiality agreements, most of it being in patent form (Plotzker et al., 1988; Hocquellet and Chenier, 1993). The type and architecture of the fibers and the microstructure of the supporting matrix determine the physical properties of C–C composites. It is, therefore, important to understand the mechanisms that control the formation of the microstructure and its interaction with the fibers in order to be able to achieve the desired composite properties at the lowest possible cost and make full use of the electrical, thermal, and mechanical properties of the fibers (Meyer, 1994). Two methods are at present used for the preparation of C–C composites: (i) liquid impregnation; and (ii) carbon vapor infiltration (CVI). A combination of the two methods is also used. Both have been described in detail in the literature (Savage, 1993a). Liquid impregnation, the topic covered in this chapter, is a two-stage method. Basically, in the first stage the fibers are impregnated with an organic precursor, giving rise to what is called a prepreg. Afterwards, the precursor is transformed into a carbonaceous matrix by thermal treatment (650–1,000 C) in an inert atmosphere via a carbonization process. Depending on the application of the material, graphitization (2,500 C) may also be required. These stages have to be repeated until the desired density is obtained (Fig. 7.1). The other method, CVI, implies the cracking of a gas and the deposition of carbon onto the fiber substrate (Chapter 7 of this book). In general, liquid impregnation is considerably quicker and cheaper than CVI processing, and the equipment required, in an ambient pressure fabrication route, is not limited © 2003 Taylor & Francis

(impregnation is carried out during winding)or dry(winding is carried out on a preimpreg- nated,dried tape).Winding can also be performed with non-impregnated fibers followed by impregnation. When using wet winding techniques,the impregnation conditions vary depending on the matrix precursor(Hirano et al.,1993),i.e.a mixture of polymers and a plasticizer,melt pitch, liquid resins,pitch,or resin dissolved in an organic solvent,or fine particles suspended in the form of an emulsion(Hocquellet and Chenier,1993).Two fundamental factors which must be controlled in this step and borne in mind when selecting the optimum conditions are:the alignment of the fibers and the homogeneous distribution of the matrix precursor.Correct fiber alignment is very important because small variations in the orientation of the fibers can mean additional stresses,which may lead to premature failure of the material.As an exam- ple,impregnation is relatively easy with solutions at low concentrations of the organic pre- cursor.However,complete removal of the solvent is hard to achieve and it may produce fiber misalignment,as well as the generation of porosity during solvent removal. The next step,after the impregnation process,involves curing and/or hot-pressing (depending on the matrix precursor),and then carbonization/graphitization.After this,other techniques(CVI,injection)are required in order to reduce the final porosity and obtain a high density material. 2.2 Pultrusion This technique is derived from wet-winding.In this case,after being impregnated in a bath, the fibers are pulled through a heated die specifically shaped to produce the desired cross- section(Fig.7.2b).Partial or complete curing occurs during the passage through the die. 2.3 Hand lay-up Fibers are placed by hand into a mould covered by a released substance and then impreg- nated with a liquid precursor by means of painting or spraying and rolling.Rolling compacts the material and ensures the effective impregnation of the fiber by the matrix precursor, homogeneous distribution and the removal of air.The fibers may be in the form of random mats or woven fabrics.Layers of impregnated fibers are stacked up(lay-up)until the desired thickness is achieved.Carbonization is then carried out to obtain a stable porous structure. Again additional techniques are needed to obtain the desired final density. 2.4 Injection technique This method(Fig.7.2c)is generally used to impregnate preforms of carbon fibers produced by dry winding or braiding and also as a way to increase the density of the composite pre- pared by other techniques.The preform is dipped into the liquefied precursor or alterna- tively the precursor is forced into the pores of the preform by means of surface tension, gravity,atmospheric pressure or external forces such as gas pressure,centrifugal force, mechanical vibration,etc.(Borokh et al.,1995).In some cases this is achieved with the help of a vacuum.Pore distribution in the preform and the rheological properties of the matrix precursor are the main factors that control this process. 2.5 Hot-press molding This technique is frequently used at laboratory scale to prepare bidirectional C-C composites (Arianoutsos et al.,1990).Woven mats of carbon fibers are stacked in a mold and sandwiched ©2003 Taylor&Francis

(impregnation is carried out during winding) or dry (winding is carried out on a preimpregnated, dried tape). Winding can also be performed with non-impregnated fibers followed by impregnation. When using wet winding techniques, the impregnation conditions vary depending on the matrix precursor (Hirano et al., 1993), i.e. a mixture of polymers and a plasticizer, melt pitch, liquid resins, pitch, or resin dissolved in an organic solvent, or fine particles suspended in the form of an emulsion (Hocquellet and Chenier, 1993). Two fundamental factors which must be controlled in this step and borne in mind when selecting the optimum conditions are: the alignment of the fibers and the homogeneous distribution of the matrix precursor. Correct fiber alignment is very important because small variations in the orientation of the fibers can mean additional stresses, which may lead to premature failure of the material. As an example, impregnation is relatively easy with solutions at low concentrations of the organic precursor. However, complete removal of the solvent is hard to achieve and it may produce fiber misalignment, as well as the generation of porosity during solvent removal. The next step, after the impregnation process, involves curing and/or hot-pressing (depending on the matrix precursor), and then carbonization/graphitization. After this, other techniques (CVI, injection) are required in order to reduce the final porosity and obtain a high density material. 2.2 Pultrusion This technique is derived from wet-winding. In this case, after being impregnated in a bath, the fibers are pulled through a heated die specifically shaped to produce the desired crosssection (Fig. 7.2b). Partial or complete curing occurs during the passage through the die. 2.3 Hand lay-up Fibers are placed by hand into a mould covered by a released substance and then impregnated with a liquid precursor by means of painting or spraying and rolling. Rolling compacts the material and ensures the effective impregnation of the fiber by the matrix precursor, homogeneous distribution and the removal of air. The fibers may be in the form of random mats or woven fabrics. Layers of impregnated fibers are stacked up (lay-up) until the desired thickness is achieved. Carbonization is then carried out to obtain a stable porous structure. Again additional techniques are needed to obtain the desired final density. 2.4 Injection technique This method (Fig. 7.2c) is generally used to impregnate preforms of carbon fibers produced by dry winding or braiding and also as a way to increase the density of the composite prepared by other techniques. The preform is dipped into the liquefied precursor or alternatively the precursor is forced into the pores of the preform by means of surface tension, gravity, atmospheric pressure or external forces such as gas pressure, centrifugal force, mechanical vibration, etc. (Borokh et al., 1995). In some cases this is achieved with the help of a vacuum. Pore distribution in the preform and the rheological properties of the matrix precursor are the main factors that control this process. 2.5 Hot-press molding This technique is frequently used at laboratory scale to prepare bidirectional C–C composites (Arianoutsos et al., 1990). Woven mats of carbon fibers are stacked in a mold and sandwiched © 2003 Taylor & Francis

by the matrix precursor powder (Figure 7.2d).The mold,matrix precursor and fibers,are heated together until the matrix precursor is thoroughly melted.At this point mechanical pressure is applied in order to place the melted precursor inside the mats and so ensure an effective impregnation.During pressurization,trapped air and excess precursor are forced out of the mold.After cooling,carbonization is required.Infiltration/carbonization cycles are repeated several times to achieve the desired density. Unidirectional C-C composites are prepared in a similar way,using a powder coating process(Klett and Edie,1995).Fiber bundles are separated by a vacuum spreader and then the individual fibers are coated with fine polymer particles in a fluidization chamber.The coated tow thus formed (tow-preg)is heated until the particles are fused to the individual filaments.Next the tow-pregs are cut,stacked and hot-pressed to form the final composite. This procedure may be specially desirable when using pitches with a high softening point and low solubility as matrix precursors. Hot-press molding is also used as part of the process of preparation of unidirectional C-C composites by wet-winding.Laminated prepregs are stacked and then hot-pressed at spe- cific temperatures,depending on the characteristics(chemical composition and rheological properties)of the precursor(Figueiras et al.,1995;Casal et al.,2001). 3 Densification efficiency The preparation of C-C composites with the right density currently requires several impregnation/carbonization cycles,as mentioned above.This is because in the initial stages of the carbonization process of the organic precursor there is a major weight loss of volatiles which continues,to a lesser extent,throughout the process as the temperature increases, giving a porous material.The pores are produced,either in the initial stages of the prepara- tion of the composite (impregnation process and molding),or during subsequent carboniza- tion.Pores of the first type are usually open (Fig.7.3a),whereas those developed on carbonization may be either open or closed(Fig.7.3a).Carbonization pores may be devel- oped either while the matrix precursor is still plastic or as a result of stresses(cracks)due to structural rearrangements or thermo-mechanical effects (Fig.7.3b).Several studies have been carried out in relation with the characterization of porosity in C-C composites (Jortner, 1986;Granda et al.,1998).All of these structural imperfections have an adverse effect on the bulk properties. A knowledge of the parameters which control the efficiency and effectiveness of densifi- cation is essential for the optimization of the process.Experience shows that the first step of impregnation is not enough to ensure that all of the empty pores in a C-C composite are filled by the matrix precursor(Fig.7.3c and d).An impregnation efficiency parameter,Y, can be defined as the relationship between the volume of impregnant introduced into the pores of the composite and the volume of the open porosity of the composite (eq.(1)).This parameter can be calculated from the increase in weight gain of the composite during the impregnation,AWi,and the volume fraction of open porosity of the composite before impregnation,0: Y= Volume ofimpregnant introducted AW p(0) Volume ofinitial empty pores Pp 0 (1) ©2003 Taylor&Francis

by the matrix precursor powder (Figure 7.2d). The mold, matrix precursor and fibers, are heated together until the matrix precursor is thoroughly melted. At this point mechanical pressure is applied in order to place the melted precursor inside the mats and so ensure an effective impregnation. During pressurization, trapped air and excess precursor are forced out of the mold. After cooling, carbonization is required. Infiltration/carbonization cycles are repeated several times to achieve the desired density. Unidirectional C–C composites are prepared in a similar way, using a powder coating process (Klett and Edie, 1995). Fiber bundles are separated by a vacuum spreader and then the individual fibers are coated with fine polymer particles in a fluidization chamber. The coated tow thus formed (tow-preg) is heated until the particles are fused to the individual filaments. Next the tow-pregs are cut, stacked and hot-pressed to form the final composite. This procedure may be specially desirable when using pitches with a high softening point and low solubility as matrix precursors. Hot-press molding is also used as part of the process of preparation of unidirectional C–C composites by wet-winding. Laminated prepregs are stacked and then hot-pressed at specific temperatures, depending on the characteristics (chemical composition and rheological properties) of the precursor (Figueiras et al., 1995; Casal et al., 2001). 3 Densification efficiency The preparation of C–C composites with the right density currently requires several impregnation/carbonization cycles, as mentioned above. This is because in the initial stages of the carbonization process of the organic precursor there is a major weight loss of volatiles which continues, to a lesser extent, throughout the process as the temperature increases, giving a porous material. The pores are produced, either in the initial stages of the preparation of the composite (impregnation process and molding), or during subsequent carbonization. Pores of the first type are usually open (Fig. 7.3a), whereas those developed on carbonization may be either open or closed (Fig. 7.3a). Carbonization pores may be developed either while the matrix precursor is still plastic or as a result of stresses (cracks) due to structural rearrangements or thermo-mechanical effects (Fig. 7.3b). Several studies have been carried out in relation with the characterization of porosity in C–C composites (Jortner, 1986; Granda et al., 1998). All of these structural imperfections have an adverse effect on the bulk properties. A knowledge of the parameters which control the efficiency and effectiveness of densification is essential for the optimization of the process. Experience shows that the first step of impregnation is not enough to ensure that all of the empty pores in a C–C composite are filled by the matrix precursor (Fig. 7.3c and d). An impregnation efficiency parameter, Yi, can be defined as the relationship between the volume of impregnant introduced into the pores of the composite and the volume of the open porosity of the composite (eq. (1)). This parameter can be calculated from the increase in weight gain of the composite during the impregnation, Wi , and the volume fraction of open porosity of the composite before impregnation, : Y (1) i Volume of impregnant introducted Volume of initial empty pores Wi p (0) © 2003 Taylor & Francis

will be lower than anticipated knowing carbon yields from the impregnant,when carbonized alone.The physico-chemical properties of the precursor,the number of pores,the structure of the pore networks in the composite (size,shape,and orientation),and the experimental conditions used are the main controlling factors. Studies of the densification efficiency of C-C composites with pitches (Oh and Park, 1994;Granda et al.,1998)have led to the conclusion that impregnation efficiency increases, as the process is repeated due to a decrease in porosity,while retention efficiency follows an inverse trend due to the changes in the shape of the pores,as densification proceeds. Overall densification efficiency depends on impregnation efficiency and the effective carbon yield of the impregnating pitch.Pitch fluidity is the main factor controlling impreg- nation efficiency.Additionally,the tortuosity of the pores also plays an important role dur- ing the impregnation process.The use of pressure during carbonization reduces the release of volatiles and pitch bloating,leading to an increase in the effective carbon yield (Granda et al.,1998).It also reduces pore closure,thereby improving densification effi- ciency.It has also been shown that closed porosity may be transformed into open porosity via graphitization (Savage,1993b). 4 Matrix precursors The matrix of a C-C composite acts as a binder by maintaining the alignment and position of the fibers,and protects them from damage.It also distributes stresses,transferring the external load to single reinforcing filaments.The structural characteristics of the matrix and their interaction with the fibers contribute significantly to the properties of the composite. These characteristics depend on the chemical composition and physical properties of the precursor,as well as on the processing conditions.As the preparation of C-C composites involves heat treatment in an inert atmosphere to transform the matrix precursor into carbon form,the chemical and rheological behavior of the matrix precursor on pyrolysis is of con- siderable importance for selecting the optimum processing conditions.Generally speaking, a good matrix precursor should have a high carbon yield while at the same time retaining its fluidity and ability to wet the fibers.Moreover,low volumetric contraction is necessary in order to avoid fiber damage and matrix shrinkage cracks.It is not easy to find a product which fulfills all of the requirements at the same time.As an example,a high carbon yield is usually associated with high organic precursor viscosity,which may impede infiltration and wetting.On the other hand,a high carbon yield may be accompanied by an adequate viscosity but also by considerable matrix bloating on carbonization.Two main types of matrix precursors are currently used in the preparation of C-C composites by liquid impreg- nation:one is based on resins and polymers and the other is based on pitches. 4.1 Resin precursors In this group,phenolic and epoxy resins are the most commonly used and both are thermoset and need curing prior to carbonization(Fig.7.4a).Apart from the above mentioned matrix requirements,it is necessary that the resin cures rapidly at low temperature(without the evo- lution of volatiles).The carbon yield in commercially available resins usually ranges from 50 to 70 wt%,depending on the type of resin and the processing conditions(Murdie et al., 1992).These yields cannot be increased by the application of pressure during carbonization. New resin precursors(Fig.7.4b)have been developed with carbon yields of up to 85 wt% (800C)(Savage,1993c),but the price of these products is extremely high and in some ©2003 Taylor&Francis

will be lower than anticipated knowing carbon yields from the impregnant, when carbonized alone. The physico-chemical properties of the precursor, the number of pores, the structure of the pore networks in the composite (size, shape, and orientation), and the experimental conditions used are the main controlling factors. Studies of the densification efficiency of C–C composites with pitches (Oh and Park, 1994; Granda et al., 1998) have led to the conclusion that impregnation efficiency increases, as the process is repeated due to a decrease in porosity, while retention efficiency follows an inverse trend due to the changes in the shape of the pores, as densification proceeds. Overall densification efficiency depends on impregnation efficiency and the effective carbon yield of the impregnating pitch. Pitch fluidity is the main factor controlling impregnation efficiency. Additionally, the tortuosity of the pores also plays an important role during the impregnation process. The use of pressure during carbonization reduces the release of volatiles and pitch bloating, leading to an increase in the effective carbon yield (Granda et al., 1998). It also reduces pore closure, thereby improving densification efficiency. It has also been shown that closed porosity may be transformed into open porosity via graphitization (Savage, 1993b). 4 Matrix precursors The matrix of a C–C composite acts as a binder by maintaining the alignment and position of the fibers, and protects them from damage. It also distributes stresses, transferring the external load to single reinforcing filaments. The structural characteristics of the matrix and their interaction with the fibers contribute significantly to the properties of the composite. These characteristics depend on the chemical composition and physical properties of the precursor, as well as on the processing conditions. As the preparation of C–C composites involves heat treatment in an inert atmosphere to transform the matrix precursor into carbon form, the chemical and rheological behavior of the matrix precursor on pyrolysis is of considerable importance for selecting the optimum processing conditions. Generally speaking, a good matrix precursor should have a high carbon yield while at the same time retaining its fluidity and ability to wet the fibers. Moreover, low volumetric contraction is necessary in order to avoid fiber damage and matrix shrinkage cracks. It is not easy to find a product which fulfills all of the requirements at the same time. As an example, a high carbon yield is usually associated with high organic precursor viscosity, which may impede infiltration and wetting. On the other hand, a high carbon yield may be accompanied by an adequate viscosity but also by considerable matrix bloating on carbonization. Two main types of matrix precursors are currently used in the preparation of C–C composites by liquid impregnation: one is based on resins and polymers and the other is based on pitches. 4.1 Resin precursors In this group, phenolic and epoxy resins are the most commonly used and both are thermoset and need curing prior to carbonization (Fig. 7.4a). Apart from the above mentioned matrix requirements, it is necessary that the resin cures rapidly at low temperature (without the evolution of volatiles). The carbon yield in commercially available resins usually ranges from 50 to 70 wt %, depending on the type of resin and the processing conditions (Murdie et al., 1992). These yields cannot be increased by the application of pressure during carbonization. New resin precursors (Fig. 7.4b) have been developed with carbon yields of up to 85 wt % (800 C) (Savage, 1993c), but the price of these products is extremely high and in some © 2003 Taylor & Francis

(a) CH2 CH CH2 Figure 7.4 Chemical structure of(a)thermosetting phenolic resin precursor;and (b)PEEK thermoplastic resin precursor. cases the preparation procedures for bulk resin production are difficult to scale up. Furthermore,the resins tend to form closed porosity on carbonization,preventing further densification. The processing of resin composites includes initial partial curing for polymerization to take place and thereby increase the carbon yield.This is followed by hot-pressing (150-300C)or curing in an autoclave to produce a rigid solid via three-dimensional cross-linking.Then pyrolysis and a series of infiltration/carbonization cycles are carried out.The processing conditions used in the preparation of the composite depend on the time/ viscosity properties of the resin.In general,the carbonization of composites to temperatures of about 800-1,000C is carried out slowly in order to prevent the rapid evolution of gases (H2O,CH4,CO,H2).The carbonization step may take many hours or even days to be com- pleted.During carbonization,the resin is converted into a glassy,isotropic carbon(shrink- age stresses in the vicinity of the fibers may cause graphitization at temperatures >2,500C). The shrinkage of the resin during curing and carbonization produces slit-shaped pores which need to be filled by a liquid matrix precursor for subsequent carbonization if the density and mechanical properties of the composite are to be improved.The fine micropore network developed during the carbonization of the resin cannot be easily reinfiltrated even when the pores are open(Rand,1993). The density of resin-based carbons is relatively low.So,when using resin,the number of densification cycles is higher than with other precursors.As an example,six cycles were required to achieve a density of 1.65 gcm-3 in order to process multidirectional cylinders with phenolic resin(Mullen and Roy,1972).Unfortunately,none of the resins available at the moment,even the more recent ones with high carbon yields,can fulfill all the require- ments of an optimum carbon matrix precursor. 4.2 Pitch precursors 4.2.1 Pitch composition Pitches offer good possibilities as carbon matrix precursors,mainly because of the wide range of pitch products with different characteristics that are available,their relatively low ©2003 Taylor&Francis

CH2 CH2 CH2 (a) O O O n (b) C OH OH cases the preparation procedures for bulk resin production are difficult to scale up. Furthermore, the resins tend to form closed porosity on carbonization, preventing further densification. The processing of resin composites includes initial partial curing for polymerization to take place and thereby increase the carbon yield. This is followed by hot-pressing (150–300 C) or curing in an autoclave to produce a rigid solid via three-dimensional cross-linking. Then pyrolysis and a series of infiltration/carbonization cycles are carried out. The processing conditions used in the preparation of the composite depend on the time/ viscosity properties of the resin. In general, the carbonization of composites to temperatures of about 800–1,000 C is carried out slowly in order to prevent the rapid evolution of gases (H2O, CH4, CO, H2). The carbonization step may take many hours or even days to be completed. During carbonization, the resin is converted into a glassy, isotropic carbon (shrinkage stresses in the vicinity of the fibers may cause graphitization at temperatures 2,500 C). The shrinkage of the resin during curing and carbonization produces slit-shaped pores which need to be filled by a liquid matrix precursor for subsequent carbonization if the density and mechanical properties of the composite are to be improved. The fine micropore network developed during the carbonization of the resin cannot be easily reinfiltrated even when the pores are open (Rand, 1993). The density of resin-based carbons is relatively low. So, when using resin, the number of densification cycles is higher than with other precursors. As an example, six cycles were required to achieve a density of 1.65 g cm3 in order to process multidirectional cylinders with phenolic resin (Mullen and Roy, 1972). Unfortunately, none of the resins available at the moment, even the more recent ones with high carbon yields, can fulfill all the requirements of an optimum carbon matrix precursor. 4.2 Pitch precursors 4.2.1 Pitch composition Pitches offer good possibilities as carbon matrix precursors, mainly because of the wide range of pitch products with different characteristics that are available, their relatively low Figure 7.4 Chemical structure of (a) thermosetting phenolic resin precursor; and (b) PEEK thermoplastic resin precursor. © 2003 Taylor & Francis

price,easy manipulation,and the possibility of achieving high carbon yields.Pitches come primarily from coal and petroleum.Both petroleum-derived and coal-derived pitches(so called coal-tar pitches)are residues from the distillation or heat treatment of petroleum frac- tions and coal-tar,respectively,and contain a large range of aromatic molecular types and sizes (Zander,1997).These pitches differ from one another in their chemical composition (aromaticity,aromatic condensation degree,molecular size,functionality,etc.),physical properties and consequently,in their pyrolysis behavior(Perez et al.,2000).This is because of differences in their origin.Figure 7.5 shows two average molecular structures for coal-tar and petroleum pitch proposed by Kershaw and Black(1993).Coal-tar pitches are more aro- matic,with a higher degree of ring condensation than petroleum pitches and they contain carbonaceous particles (<1 um)called primary Quinoline Insolubles (QD).These particles are generated during the coke oven operation and are entrained by gases to form part of the tar.Consequently they remain in the pitch which is obtained by tar distillation.Two types of commercial coal-tar pitches are produced mainly for the aluminum and steel industry: binder pitches and impregnating pitches.These pitches differ in their specifications(soften- ing point,carbon yield,QI content,toluene insoluble content,etc.)depending mainly on the preparation process used (vacuum distillation,thermal treatment,etc.).However,the composition of the tar will also have some influence. In general,pitches are extremely complex in composition.They are composed of several hundreds and even thousands of aromatic compounds which are present in different con- centrations and differ in molecular size and topology.These compounds cover a very wide Coal-tar pitch 1 in 4 molecules has a CH2-CH3 group Petroleum pitch 1 in 4 molecules contains S atoms CH2-CHa CH3 CH CH3 Figure 7.5 Average molecular structures for coal-tar and petroleum pitch (Kershaw and Black,1993). ©2003 Taylor&Francis

price, easy manipulation, and the possibility of achieving high carbon yields. Pitches come primarily from coal and petroleum. Both petroleum-derived and coal-derived pitches (so called coal-tar pitches) are residues from the distillation or heat treatment of petroleum fractions and coal-tar, respectively, and contain a large range of aromatic molecular types and sizes (Zander, 1997). These pitches differ from one another in their chemical composition (aromaticity, aromatic condensation degree, molecular size, functionality, etc.), physical properties and consequently, in their pyrolysis behavior (Pérez et al., 2000). This is because of differences in their origin. Figure 7.5 shows two average molecular structures for coal-tar and petroleum pitch proposed by Kershaw and Black (1993). Coal-tar pitches are more aromatic, with a higher degree of ring condensation than petroleum pitches and they contain carbonaceous particles (1m) called primary Quinoline Insolubles (QI). These particles are generated during the coke oven operation and are entrained by gases to form part of the tar. Consequently they remain in the pitch which is obtained by tar distillation. Two types of commercial coal-tar pitches are produced mainly for the aluminum and steel industry: binder pitches and impregnating pitches. These pitches differ in their specifications (softening point, carbon yield, QI content, toluene insoluble content, etc.) depending mainly on the preparation process used (vacuum distillation, thermal treatment, etc.). However, the composition of the tar will also have some influence. In general, pitches are extremely complex in composition. They are composed of several hundreds and even thousands of aromatic compounds which are present in different concentrations and differ in molecular size and topology. These compounds cover a very wide Figure 7.5 Average molecular structures for coal-tar and petroleum pitch (Kershaw and Black, 1993). H H N CH3 Coal-tar pitch CH3 CH3 CH3 CH3 Petroleum pitch 1 in 4 molecules contains S atoms 1 in 4 molecules has a CH2CH3 group CH2CH3 CH2CH3 © 2003 Taylor & Francis

20 15 10 5 0 0 500 1,000 1,5002,000 2,500 3,000 Molecular weight of fraction(amu) Figure 7.6 Molecular weight distribution of the pyridine-soluble part of coal-tar pitch(Boenigk et al., 1990). range of molecular weight from ~200 to >3,000 amu (Zander,1997).As an example, Fig.7.6 shows the molecular weight distribution determined by preparative size exclusion chromatography (SEC)of the pyridine-soluble fraction (95 vol%)of a coal-tar pitch (Boenigk et al.,1990).From results obtained by preparative SEC and other techniques, these authors conclude that a typical coal-tar pitch consists of~40 wt%of compounds with a molecular mass of s330 amu,~50 wt%of compounds with a molecular mass ~330-1,500 amu,while 10 wt%of the remaining pitch consists of high-molecular weight material (between ~1,500 and >3,500 amu). Coal-tar pitches are mainly composed of polycyclic aromatic hydrocarbons(PAH)and to a lesser extent of polycyclic aromatic compounds containing alkylsubstituents and/or het- eroatoms(O,S,and N).These compounds can be compiled into categories by fractionation by means of extrography (Granda et al.,1990).This technique is able to distinguish between pitches differing in origin (coal-tar or petroleum),their uses (binding or impreg- nating)and also the preparation process (thermal treatment,vacuum distillation,etc.). Thus impregnating and binder pitches mainly differ in the distribution of PAH and the amount of basic polycyclic aromatic compounds(Bermejo et al.,1994).A typical impregnating pitch has a higher content in both PAH of lower molecular weight and basic nitrogen compounds. There is also what is considered to be a third generation of pitches.Some of these are commercially available(Mitsubishi Gas Chemical and Mitsubishi Oil)and others are still under development.This third group includes the mesophase pitches(from organic precur- sors and from the pyrolysis of commercial pitches)and isotropic pitches of high carbon yield(from the direct treatment of tar,commercial pitches,and other tar distillation fractions such as anthracene oil).These will be discussed in Section 5 of this chapter. 2003 Taylor Francis

range of molecular weight from ~200 to 3,000 amu (Zander, 1997). As an example, Fig. 7.6 shows the molecular weight distribution determined by preparative size exclusion chromatography (SEC) of the pyridine-soluble fraction (95 vol %) of a coal-tar pitch (Boenigk et al., 1990). From results obtained by preparative SEC and other techniques, these authors conclude that a typical coal-tar pitch consists of ~40 wt % of compounds with a molecular mass of 330 amu, ~50 wt % of compounds with a molecular mass ~330–1,500 amu, while 10 wt % of the remaining pitch consists of high-molecular weight material (between ~1,500 and 3,500 amu). Coal-tar pitches are mainly composed of polycyclic aromatic hydrocarbons (PAH) and to a lesser extent of polycyclic aromatic compounds containing alkylsubstituents and/or heteroatoms (O, S, and N). These compounds can be compiled into categories by fractionation by means of extrography (Granda et al., 1990). This technique is able to distinguish between pitches differing in origin (coal-tar or petroleum), their uses (binding or impregnating) and also the preparation process (thermal treatment, vacuum distillation, etc.). Thus impregnating and binder pitches mainly differ in the distribution of PAH and the amount of basic polycyclic aromatic compounds (Bermejo et al., 1994). A typical impregnating pitch has a higher content in both PAH of lower molecular weight and basic nitrogen compounds. There is also what is considered to be a third generation of pitches. Some of these are commercially available (Mitsubishi Gas Chemical and Mitsubishi Oil) and others are still under development. This third group includes the mesophase pitches (from organic precursors and from the pyrolysis of commercial pitches) and isotropic pitches of high carbon yield (from the direct treatment of tar, commercial pitches, and other tar distillation fractions such as anthracene oil). These will be discussed in Section 5 of this chapter. 20 15 Weight fraction (wt %) 10 5 0 0 500 1,000 Molecular weight of fraction (amu) 1,500 2,000 2,500 3,000 Figure 7.6 Molecular weight distribution of the pyridine-soluble part of coal-tar pitch (Boenigk et al., 1990). © 2003 Taylor & Francis