Carbon fiber HS from PAN [UIDCarbFibHS]

Reviewed by Michael Overcash 5-17-09 Rereviewed by Michael Overcash 5-25-09 Final Carbon HS from PAN [UIDCarbFibHS] CONTENTS OF FACTORY GATE TO FACTORY GATE LIFE CYCLE INVENTORY SUMMARY Chemistry... 2 Process Summary... 3 Summary of LCI Information...14 Process Diagram Interpretation Sheet...16 Process Diagram or Boundary of LCI...17 Mass Balance of Chemicals in Each Process Stream...19 Graph of Cumulative Chemical Losses through Manufacturing Process...27 Graph of Cumulative Contaminated Water Use / Emission through Manufacturing Process...28 Graph of Cumulative Non-Contaminated Water Use / Emission through Manufacturing Process...29 Energy Input for each Unit Process, Cumulative Energy Requirements, Cooling Requirements (exotherms), and Assumed Heat Recovery from Hot Streams Receiving Cooling...30 Graph of Cumulative Energy Requirements...32 Authors E. Griffing Peer reviews, name (date) MR Overcash (5-25-2009) Gtg report last modified on 10-4-2010 Additional notes Checked for database consistency on 10-4-2010 First gtg version finalized on 5-25-2009 Modification history, Author (date) EMG (10-4-2010) and EMG (5-25-2009) Products Standard inputs carbon, high strength PAN precursor, Oxygen from air, Nitrogen Methodology: Environmental Clarity gtg lci reports are based on industrial practice information, standard methods of engineering process design, and technical reviews. These reports are intended to be representative of industrial production based on the stated route. Terms of use: Environmental Clarity does not assume any liability due to use of these lci data. Integration of these data with lci data based on other methodologies is the responsibility of the user. Each report may be updated to improve model accuracy or representativeness. Users of this report should cite: E. Griffing and M. Overcash, Chemical Life Cycle Database, www.environmentalclarity.com, 1999 - present. updated on 10/4/2010 1

Chemistry Primary reaction: (looks like this needs updating) PAN precursor + O 2 stabilized PAN+ H 2 O + CO 2 + CO + HCN (1)* stabilized PAN H 2 O+carbon s + CO 2 +CO+HCN+N 2 +NH 3 +C2H 6 +CH 4 +C 2 H 4 (2) Notes: * details are specified in the literature section Combustion reactions: (type of side rxns) CO+1/2 O 2 CO 2 (5) 2HCN+4.5O 2 2NO 2 + 2CO 2 + H 2 O (6) 2HCN+2.5O 2 N 2 +2CO 2 +H 2 O (7) H 2 +0.5O 2 H 2 O (8) 2NH 3 +1.5O 2 N 2 +3H 2 O (9) C 2 H 6 + C 2 H 4 +6.5O 2 4CO 2 +5H 2 O (10) CH 4 +2O 2 CO 2 +2H 2 O (11) updated on 10/4/2010 2

Process Summary Literature Carbon s for textile uses are typically made from pitch, rayon, or polyacrylonitrile (PAN). Fibers used in critical structural applications are typically made from PAN. Carbon s made from mesophase pitch are graphitic, have a high Young's modulus, and high thermal conductivity. Fibers made from PAN are turbostatic and have a higher tensile strength (Wikipedia, 2009). PAN s for the aerospace industry are manufactured by Toray Industries, Toho Tenax, Mitsubishi Rayon, Hexcel, Cytec Industries, and Schunk Gruppe. Commercial grade PAN s are made by Zoltek, SGL Carbon, and Fortafil (Wikipedia, 2009). The carbon content of acrylonitrile is 68%, and the carbon yield in the process is 50-55% (Morgan, 2005). Typically, commercial PAN s are made from between 85 and 98% acrylonitrile with the balance from comonomers. Issues relating to the composition of the PAN precursor are discussed in the PAN GTG report. In this GTG report, we use PAN based s as an input. The manufacture of carbon s takes place in four phases: oxidation, carbonization, surface treatment, and sizing (Kelly, 2009; Wikipedia, 2009; Morgan, 2005). Most of the non-carbon atoms are removed through volatilization. High strength s contain 92-95% carbon and about 5 % nitrogen. During oxidation (stabilization), the PAN is converted into a cyclic structure. In addition to cyclization, the oxidation process yields three-dimensional cross-linking by oxygen bonds. The oxygen bonds take the form of OH, C=O, and -COOH. An oxygen content of fully stabilized s is 6-8% (Morgan, 2005) or 8 to 12 wt% (Cheung, 1994; Morgan, 2005). While oxygen is being incorporated into the s, HCN, ammonia and other gases evolve, and the overall weight change is small (Cheung, 1994; Morgan, 2005). During carbonization, the s lose about 40% by weight due to volatilization of HCN, NH 3, H 2, CO 2, and CO. From an energy perspective, the important factors in the process are the utilities required in the carbonization furnaces, which are typically heated by electricity, and the loss of material during carbonization. The unit temperatures and yield data are given Table 1. According to Kelly (2009), carbon used in wind turbines has a modulus of around 230, which is in the high range of standard modulus. Graphitization produces high modulus s (modulus of 325 GPa or higher). These high-modulus s are not used for high strength applications, and are excluded from Table 1. Table 1. Reactor conditions and yields for intermediate modulus s. Oxidation temperature, oc Carbonization temperature, oc Cheung (1994) 180-300 300-1200, type III 1000-1700, type II 1500-2800, type III Kelly (2009) 300 1000, low modulus 1000-1500, standard modulus 1500-2000, intermediate modulus 2000+, high modulus Ullmann's (2005) 200-300 1200-1500, high strength Morgan (2005) 220-270 Up to 1500 50-55 CEH (2007) 190-280 1000-1300 Gupta et al. (1991) Kirk Othmer (2007) 200-300 1000-1700 Minus and Kumar 200-300 1000-1700 1500-1600, maximum strength Used in this GTG 300 1300 55 report Overall yield, wt % of input 62% based on preoxidized wt. > 50% carbon yield 50-60% updated on 10/4/2010 3

Oxidation reactor During the oxidation process, the PAN undergoes cyclization and oxidation reactions, and both of these reactions are exothermic. A simple model of these two reactions is given by Kirk-Othmer (2007) as shown in Figure 1. (1) Figure 1. Cyclization and oxidation reactions from Kirk Othmer (2007) Additional structures are given by Gupta et al. (1995) as shown in Figure 2. The percentage values show the approximate composition of the oxidized PAN s in mole percent of the groups in the PAN. The structure labeled (c) is formed in reaction 1 from Kirk-Othmer. From structure c, a simple dehydration reaction produces the structure labeled (b) or the oxidation reaction (2) produces structure a. Structures a through c and the mole percents are based on work by Morita et al. (1986) (citation in Morgan, page 249). Estimates for the heats of reaction to form structures a through d are given by Gupta et al. (1995). These estimates are used to calculate an overall heat of reaction per kg of PAN reacting. (2) (a) (b) (c) (d) Figure 2. Structures formed in the oxidation stage (Gupta et al., 1995) updated on 10/4/2010 4

Table 2. Heat of reaction in oxidation reactions mol fraction heat of reaction for component, MJ/kgmol component portion of total heat of reaction, kj/gmol total mass of structure, kg/kgmol portion of total mass, kg / kgmol total heat of reaction, MJ/kg a 0.4-303.6-121.44 91 36.4 b 0.3-118.2-35.46 75 22.5 c 0.2-94 -18.8 79 15.8 d 0.1-517.5-51.75 217 21.7 Total -227.45 96.4-2.36 The off-gas from the oxidation reactions contain hydrogen cyanide, water, ammonia, carbon monoxide, and carbon dioxide. According to Gupta et al. (1991; 1995), chain scission reactions lead to formation of HCN, CO, and CO 2. Morgan (2005) gives additional reactions for the production of HCN are shown in Figure 3. The quantity of each of the evolved gases is given in Table 3. The heats of reaction from the formation of these chemicals is excluded from the model. The oxygen uptake based on the structures in Table 3 is about 10%, which is consistent with the literature values. However, the material from the PAN that is evolved is only 4 g/100g PAN (Table 3). According to Morgan (2005), the overall weight loss during oxidation is 3 wt%. Thus, we would expect about 11% of the PAN material is evolved. This inconsistency indicates that the off-gases given by Morgan (2005) in Table 3 may be too low. In this GTG report, we scale the off-gas evolution up to produce an overall loss of 11 wt% (of the PAN input) 8 wt% O and 3 wt% loss. Figure 3. Additional HCN producing reactions, Morgan (2005) updated on 10/4/2010 5

Table 3. Off-gas from the oxidation reactions. Data from Morgan (2005) Used in this GTG report Amount Amount evolved, Amount evolved, Amount evolved, evolved, g/100g PAN Amount evolved, g scaled to 11 wt%, g scaled to 11 g/100g carbon (assume 55 wt% non-oxygen/100 g non-oxygen / 100 g wt%, g / 100 g yield) PAN PAN PAN HCN 2.16 1.19 1.19 3.39 3.39 H2O 19.6 10.8 1.20 3.42 30.8 CO2 7.50 4.13 1.13 3.21 11.8 CO 1.00 0.550 0.236 0.673 1.57 NH3 0.190 0.105 0.105 0.298 0.298 3.85 11.0 47.8 Carbonization reactors The carbonization typically takes place in a low temperature reactor (up to 1000 oc), a high temperature reactor (up to 1500 oc), and a high modulus furnace for higher temperatures if needed (Morgan, 2005). During carbonization, the primary volatile byproducts are HCN, NH 3, H 2 O, CO 2, and N 2, and hydrocarbons. The final product after carbonization at around 1300 o C contains mostly carbon (92-95%) and up to 5% nitrogen. Sample reactions during carbonization are given by Gupta (1991) and reproduced in Figure 4. The full set of reactions are more complicated and not fully understood. At lower temperatures, CO, CO 2, water, and HCN are evolved. At higher temperatures, mostly N 2 and H 2 are evolved (Gupta et al., 1991). updated on 10/4/2010 6

Figure 4. Reactions during carbonization (Gupta, 1991). The quantity of each off-gas was reported by Cheung (1994) as shown in Figure 5. The mix of offgases depends on the amount of oxygen that was incorporated during stabilization. The balance of CH 4, CO, and CO 2 shifts toward CO 2 when more oxygen if available. The amount of H 2 and H 2 O shifts in a similar manner as more oxygen is available. If we use the emissions for 12% O, which is closest to the 8% used in our model, the total non-oxygen emissions are 0.27 g / g. These emissions are lower than expected based on the carbon yield, and the amounts of C, H, and N in the emissions are inconsistent with the amounts of these atoms in the input. In order to develop a model that is self-consistent, we use a mass balance on each element. We also make some assumptions on the emissions to make them as consistent as possible with the empirical data. updated on 10/4/2010 7

Figure 5. Emissions during pyrolysis of PAN based pre-oxidized precursor (Cheung, 1994) Overall mass balance The input is typically over 95% PAN with some comonomers. We base the overall mass balance on 100% PAN, which has a repeating structure with the formula C 3 H 3 N. We also assume that the carbon product is 95% C and 5% N. The mass of carbon, nitrogen, and hydrogen in the input PAN, final product, and evolved gases are given in Table 4. Excluding the higher hydrocarbons, there are a total of ten emissions (HCN, NH 3, N 2, CH 4, CO, CO 2, C 2 H 4, C 2 H 6, H 2, and H 2 O). From the mass balance, we have four equations (one specifying the amount of each element that is evolved, reactions a - d in Figure 6) and ten unknowns. Thus, we need some empirical data to specify the emissions. The molar ratio of CH 4 to CO to CO 2 is assumed to be 2:1:1. In addition, the molar ratio of CH 4 to C 2 H 4 to C 2 H 6 is assumed to be 10:1:1. These ratios are consistent with Figure 5. Thus, we have four additional equations (h-k). The percent loss in low temperature carbonization was given as 30% and that of high temperature carbonization was given as 15% (Morgan, 2005). We had a weight loss of 3% in oxidation; therefore, we used 28% and 14% for losses in the low and high temperature carbonization stages. The low temperature volatilization is primarily due to CH 4,CO,CO 2, water, and HCN and the high temperature volatilization is primarily due to N 2 and H 2 (Gupta et al., 1991). In this GTG report, we assume that all of the low temperature volatilization is due to CH 4, CO, CO 2, C 2 H 4, C 2 H 6, water, and HCN, and all of the high temperature volatilization is due to N 2 and H 2. This allows us to write two additional equations (e and f). However, these are not independent equations, because the sum of equations a through d and the sum of equations e and f are identical. Thus, we omit equation e, and one more equation is needed. We assume that the molar emissions of HCN are equal to those of N 2 (equation g), which is consistent with the emissions data from Figure 5 for low oxygen values. The emissions from our model are compared to those given by Cheung (1994) in Figure 7. Our total emissions are higher, which is consistent with the overall mass yield. In addition, our net emissions of updated on 10/4/2010 8

g/g oxidized PAN nitrogen are much higher due to the quantity of hydrogen and nitrogen in the input s. The emissions calculated from this model are shown in Table 5. Table 4. Mass balance on atoms in PAN, carbon (CF), and evolved gases in oxidation furnace. PAN, wt% stabilized PAN, wt% CF, wt% PAN mass, kg stabilized PAN mass, kg CF Mass evolved in mass, kg carbonization, kg C 67.9 64.5 93.0 123 114 95.0 18.7 N 26.4 25.2 5.00 48.0 44.4 5.00 39.4 H 5.7 2.1 2 10.3 3.7.00 3.7 O 8.0 14.5 14.5 Sum 100 100 100 181.8 176.4 100 76.4 (12/16)CH 4 +(12/28)CO+(12/44)CO 2 +(24/28)C 2 H 4 +(24/30)C 2 H 6 +(12/27)HCN=18.7 C balance (a) (14/27)HCN + (14/17)NH 3 + 1N 2 = 39.4 N balance (b) (4/16)CH 4 +(4/28)C 2 H 4 +(6/30)C 2 H 6 +(2/18)H 2 O + (1/27)HCN+(3/17)NH 3 +1 H 2 =3.7 H balance (c) (16/28)CO+(32/44)CO 2 +(16/18)H 2 O = 14.5 O balance (d) 1CO + 1H 2 O + 1HCN + 1NH 3 = 50.9 evolved in LT carbonization (e) 1N 2 + 1H 2 = 25.5 evolved in HT carbonization (f) -1/27HCN + 1/28N 2 = 0 (g) -1/16CH 4 + 2/28CO = 0 (h) -1/16CH 4 + 2/44CO 2 = 0 (i) -1/16CH 4 + 10/28C 2 H 4 = 0 (j) -1/16CH 4 + 10/30C 2 H 6 = 0 (k) Figure 6. Equations used in mass balance model. 0.16 0.14 Cheung (1994) mass balance 0.12 0.1 0.08 0.06 0.04 0.02 0 H2 N2 CO CH4 CO2 NH3 H2O HCN C2H4 C2H6 carbonization emissions Figure 7. Emissions from carbonization of PAN s. updated on 10/4/2010 9

Table 5. Emissions from carbonization of PAN s. g/100g stabilized PAN H 2 0.013 N 2 13.8 CO 2.06 CH 4 2.36 CO 2 3.24 NH 3 1.317 H 2 O 5.29 HCN 14.3 C 2 H 4 0.412 C 2 H 6 0.442 Combustion of off-gases Off-gases from the oxidation process are reacted at 300 o C with a platinum group metal catalyst (Morgan, 2005). These oxidizers remove HCN and NH 3 with a 95% efficiency. A more recent proprietary thermal oxidizer can recover heat and is 99% efficient. We base this report on a temperature of 300 o C and 95% efficiency. Both the low and high temperature ovens are purged with N 2 to prevent oxygen from entering the oven. Common practice is to directly incinerate the effluent gases with a gas or oil fired incinerator. When the oven reaches higher temperatures, the off-gases become a significant source of fuel, and the supplemental gas fuel requirement is reduced. The flow of air is set to give an operating temperature of 750 o C in the incinerator (Morgan, 2005). HCN removal efficiencies can be as high as 99%, but they are typically 90-95% efficient (Morgan, 2005). A two-stage combustion process can be employed. In this case, the first reactor provides a reducing environment, breaking larger molecules into H 2, H 2 O, CO, and CO 2, and converting NOx into N 2. The effluent from the reduction furnace is quenched to 750 o C, and the H 2 and CO is oxidized at 750 oc (Morgan, 2005). The effluent from the thermal oxidizer can then be used to produce steam in a heat recovery boiler. In this GTG model, we assume a temperature of 750 o C is required for oxidation. The gas effluent from the reactor is cooled to from 1300 o C to 750 o C prior to the incinerator. Nitrogen oxide emissions in typical low-nox burners are around 100 ppm (US7497942; US 5215455; US 5240690). We assume an NO 2 content of 100 ppm in the effluent gas of the combustion reactors. Energy balance in carbonization and oxidation Morgan (2005) gives an example of thermodynamic calculations for a carbonization reactor. In addition to specific heats, there are losses to the surrounding environment and losses in the electrical supply. The flow of N 2, which is used to remove evolved gases also affects the energy requirement. The important parameters from the example calculations are reproduced in Table 6. The energy calculations are based on a final temperature of 1400 o C. No heat is associated with reactions in Morgan's model. The overall thermal efficiency of the reactor is calculated as the specific heat required / total electricity used or 69%. In this report, we base the energy input to the process on our thermodynamic model, and use a thermal efficiency of 69% to determine the total power consumption. updated on 10/4/2010 10

Table 6. Parameters of carbonization reactor N 2 flow, kg/kg CF 0.646 Heat carbon, MJ/kg CF 2.15 Heat N 2, MJ/kg CF 0.94 Heat flux to surroundings, MJ/kg CF 0.228 Conductive loss, power feeder, MJ/kg CF 0.669 Resistive loss, power feeder, MJ/kg CF 0.486 Total power, MJ/kg CF 4.4714 In the oxidation reactor, the required airflow was calculated to ensure that the HCN concentration does not exceed the explosive limit of 5 volume%. We also apply an engineering safety factor and use an HCN content of 2.5% in the effluent. Using an HCN production of 0.034 g/g PAN gives an air flow of 1.46 g/g PAN. Morgan states that the volume of gas produced in the oxidation reactor is 10 5 times the volume of produced. This air flow would require 78 g air / g. We believe that this is incorrect, and use an air flow that provides twice the stoichiometric requirement of oxygen uptake in the. This gives 3.8 g air / g PAN into the process. We assume that the electrical losses (MJ loss / MJ delivered) are the same in the oxidation reactor as in the carbonization reactor (26%). However, due to much lower temperature of the reactor, we do not show any heat flux to the surroundings. Surface treatment Surface treatment is done prior to sizing in order to make the sizing material stick better. After treatment, the is washed and dried to 0.1% moisture (Morgan, 2005). A typical treatment is an electrolytic oxidation (Morgan, 2005; US 4,839,006). The is run through an aqueous electrolyte, such as ammonium bicarbonate, and a charge is applied, such that the is the anode (Morgan, 2005). In some cases, a two step treatment is used, where the first electrolyte may be an inorganic acid (phosphoric, sulfuric, or nitric), and the second treatment may be an ammonium carbonate or similar with ph of 7 or higher. In this GTG report, we show a single electrolytic step using sulfuric acid followed by a single wash. After passing through the electrolyte, the carbon carries about 40 g electrolyte solution / 100 g carbon to the wash (US 5,078,840). The drying is a vacuum mechanical water removal, estimated to yield 10 g electrolyte solution/100 g CF and then thermal drying. After the wash and dry, the has about 150 ppm electrolyte (US 4,401,533). If 40 g of wash are entrained per 100 g of during the wash, and the wash is continuous the electrolyte concentration in the wash is 0.04% electrolyte. This corresponds to a water requirement of 50,000 kg per 1000 kg of untreated. We assume that some efforts are taken to reduce water consumption, such as multiple washes or countercurrent wash design, and show only 2000 kg of water used in the wash (per 1000 kg of untreated carbon ). Surface treatment electricity consumption Surface treatment requires a small electrical input. According to US 4,839,006, a typical example uses 125 coulomb/g with a voltage of 0.2 to 0.8. Using 0.4 volts, gives an energy consumption of 50 MJ/1000kg. US 4,704,196 gives a current of 5-20 amperes/m 2, a voltage of 3-15 V, and a speed of 0.5 to 2 m/minute and a path of 1.1 m in the bath. Using typical figures in example 1 of 4,704,196 (5V, 8 amperes/m 2, one minute, gives 40 amp-vol-min/m 2. A diameter of 6 microns (Gupta et al., 1991) and a density of 1.6 g/cm 3, we calculate an energy consumption of 1000 MJ/1000kg. US 4,401,533 gives the electric requirement as 0.02 to 8 V*A*min / m 2, and values in example one were 0.4 to 2 V*A*min/m2. Using two V*A*min/m 2, a diameter of 6 m, and density of 1.6 g/cm 3, we calculate an energy consumption of 50 MJ/1000kg. Morgan (2005) does not give a typical electricity requirement, but states that a high modulus needs about 20 times the treatment level as a high strength. We use an estimate of 50 MJ/1000kg for surface-treatment. updated on 10/4/2010 11

Sizing Carbon s come into contact with guides and rollers during processing. To protect the surface during these operations, a 'size' is added to the surface. The size can be solvent-based or an aqueous emulsion of epoxy resin or other polymer, and the choice depends on the resin to be used, because the sizing also forms a surface for bonding with the resin (Cheung, 1994). US 5,685,554 describes applying a water-based emulsion of one ester and one polyepoxy compound. After drying at 120 o C, the sizing agent is applied at a mass of 0.1 to 5 percent of the carbon weight (US 5,688,544; US 5,334,419) or 0.5 to 7% by weight (US 5,063,261). In US 4,923,752, a 9:9:1 solution of ethanol:nmp:solids was used in sizing, and the carbon picked up 1 g solution per g, yielding a solid sizing addition of 1 wt%. Drying was at 170 o C. In this model, we assume 1 g of size is added per 100 g of. The size is added as a 2% emulsion in water. Thus, 50 g of water/size emulsion per 100 g of are picked up as the passes through. The excess water is removed in a dryer at 170 o C. LCI design Rolls of PAN precursor s are placed on creels, and wound through an oxidation reactor. Tension is applied to the s as they pass through the reactor at 300 o C. Energy from the motors driving and tensioning the are not included in the model. Air is blown through the reactor to provide twice the stoichiometric quantity of oxygen required, which corresponds to 3.8 g air per g PAN into the reactor. Exhaust from the oxidation furnace is oxidized at 300 o C in a catalyzed incinerator, which reduces HCN and CO content by 95%. Exhaust from the incinerator is cooled to 25 oc with heat recovery. The stabilized s are fed to a carbonization furnace. We show a single furnace as opposed to a separate low and high temperature furnace. The furnace temperature is 1300 o C, and the off-gases exit at 1300 o C. The off-gases are mixed with air and incinerated at 750 o C. At steady state, this reactor requires cooling, and the exit gases from the incinerator are cooled to 25 o C with heat recovery. The carbonized s are cooled to 60 oc with ambient cooling, and passed through a surface treatment bath, also at 60 oc. The surface treatment utilized an ammonium carbonate solution bath and electric current. The electrolyte coats the at 40 g solution per 100 g. A makeup stream is used to replace the electrolyte, and the entrained solution is washed off in a bath. The pre-treated s are then dried to 0.1% moisture at 100 oc, and cooled to 60 oc prior to sizing. During sizing, 50 g of a sizing emulsion are entrained onto the. The solids content in emulsion is 2%, giving a size of 1g / 100 g of carbon. The sized s are then dried at 170 oc and cooled to 25 oc. The process has a basis of 1000 kg of dry weight, not including the size. Thus, the dry weight is 1010kg. References: CEH (2007) Chemical Economics Handbook Marketing Research Report 542.4000 A. Carbon Fibers, (Ngan Tefera, Thomas Kalin, and Kazuteru Yokose, article authors). Cheung, D.D.L. (1994) Carbon Fiber Composites, Butterworth-Heinemann. Grassie, N., McGuelun R. (1972). Pyrolysis of polyacrylonitrile and related polymers, European Polymer Journal, 8(2), pp. 257-269. Gupta, A.K., Paliwal, D.K., Bajaj, P. (1991) Acrylic Precursors for Carbon Fibers, Polymer Reviews, 31 (1), pp 1-89. Gupta, A.K., Paliwal, D.K., Bajaj, P. (1993) Acrylonitrile-Acrylic Acids Copolymers. 1. Synthesis and Characterization, Journal of Applied Polymer Science, 49, pp 823-833. Gupta, A.K., Paliwal, D.K., Bajaj, P. (1995) Effect of an Acidic Comonomer on Thermooxidativ Stabilization of Polyacrylonitrile, Journal of Applied Polymer Science, (58), pp 1161-1174. Gupta, A.K., Paliwal, D.K., Bajaj, P. (1998) Melting Behavior of Acrylonitrile Polymers, Journal of Applied Polymer Science, (70), pp 2703-2709. Kirk Othmer (2007) Kirk Othmer Encyclopedia of Chemical Technology, Carbon Fibers (M.L. Minus and S. Kumar, article authors). Kelly, V. (2009) online document: url: http://www.geocities.com/vpkelly.geo/ Morgan, P. (2005) Carbon Fibers and their Composites, CRC Press, Taylor and Francis Group. updated on 10/4/2010 12

Tsai, J.S. and Lin C. H. (1991) Effect of Comonomer Composition on the Properties of Polyacrylonitrile Precursor and Resulting Carbon Fiber, Journal of Applied Polymer Science, 43(4) pp 679-685. Ullmann's (2005) Ullmann's Encyclopedia of Industrial Chemistry, Composite Materials, (Ilschner, B., Lees, J.K., Dhingra, A.K., McCullough, R.L, article authors). US 5,334,419 (1994) Method of Sizing Carbon Fibers, Aichi, Japan. Wikipedia (2009) Carbon. http://en.wikipedia.org/wiki/carbon_ Critical parameters / Yield information from both reactors Total yield of Process: (% yield produced by the overall process based on reactant input to process) Notes: From mass balance of or Yield from PAN precursor 55% Product purity carbon, Comments Used here 95% C, 5% N Additional 1 g size added per 100 g carbon. updated on 10/4/2010 13

Summary of LCI Information Standard inputs UID Name Flow Purity Units Comments UIDPANPCF PAN precursor 1816 - [kg/hr] UIDO2FromAir Oxygen from 1540 - [kg/hr] air 7727-37-9 Nitrogen 940 - [kg/hr] Total 4296 [kg/hr] Non-reacting inputs UID Name Flow Purity Units Comments 7732-18-5 Water 2880 - [kg/hr] UIDO2FromAir Oxygen from 1219 - [kg/hr] air 7727-37-9 Nitrogen 9080 - [kg/hr] Total 1.32E+04 [kg/hr] Ancillary inputs UID Name Flow Purity Units Comments UIDSizingSolids sizing solids 10.0 - [kg/hr] 7664-93-9 Sulfuric acid 20.0 - [kg/hr] Total 30.0 [kg/hr] Products UID Name Flow Purity Units Comments UIDCarbFibHS carbon, high strength 1000 100 [kg/hr] based on carbon portion, excludes size material Total 1000 [kg/hr] Benign outflows UID Name Flow Purity Units Comments UIDSizingSolids sizing solids 10.0 - [kg/hr] part of product 7732-18-5 Water 3786 - [kg/hr] 7782-44-7 Oxygen 1219 - [kg/hr] 7727-37-9 Nitrogen 1.04E+04 - [kg/hr] Total 1.55E+04 [kg/hr] Process emissions UID Name Gas Liquid Solid Solvent Units Comments 7664-93-9 Sulfuric 0 19.9 0 0 [kg/hr] acid 74-84-0 Ethane 0.0101 0 0 0 [kg/hr] 7664-41-7 Ammonia 1.16 0 0 0 [kg/hr] 74-90-8 Hydrogen 15.7 0 0 0 [kg/hr] cyanide 630-08-0 Carbon 3.24 0 0 0 [kg/hr] monoxide 124-38-9 Carbon 1013 0 0 0 [kg/hr] dioxide Total 1034 19.9 0 0 updated on 10/4/2010 14

Mass balance Total inputs 1.75E+04 Total outflows 1.75E+04 Net input 1.54 Energy use Energy type Amount Units Comments electricity 6986 [MJ/hr] Net electricity use at plant heating steam 3099 [MJ/hr] heating by steam (0.85 efficiency included) Net input requirement 1.01E+04 [MJ/hr] Net of energies input to system cooling water - 1.95E+04 [MJ/hr] net cooling by cooling water potential recovery - 1.07E+04 [MJ/hr] potential energy recovery (negative) Net energy -646 [MJ/hr] Net input requirement - potential recovery updated on 10/4/2010 15

Process Diagram Interpretation Sheet 1) As much as possible, standard symbols are used for all unit processes. 2) Only overall input and output chemicals are labeled on these diagrams. All intermediate information is given on the attached Process Mass Balance sheet 3) The physical state of most streams is shown (gas, g; liquid, l; solid, s) 4) The process numbering is as follows, generally numbers progress from the start to the end of the process numbers are used for process streams C i, i = 1,..n are used for all cooling non-contact streams S j, j = 1,...n are used for all steam heating non-contact streams 5) Recycle streams are shown with dotted lines For most streams, the temperature and pressure are shown, if the pressures are greater than 1 atm updated on 10/4/2010 16

Process Diagram or Boundary of LCI Steam enters the process as a gas at 207 o C and leaves as a liquid at 207 o C. Cooling water enters at 20 o C and leaves at 50 o C. Unless otherwise indicated, all processes are at 1 atm and 25 o C. Fugitive Losses (Total) (g) 0 kg/hr 4 (g) 300 o C C1 C2 C3 5 (g) 300 o C HX 1 C4 6 (g) 5358 kg Nitrogen 707 kg Oxygen 579 kg Water 352 kg Carbon dioxide 3.08 kg Hydrogen cyanide 1.43 kg Carbon monoxide 0.700 kg Nitrogen dioxide 25.0 o C air PAN s Off-wind creels 1 (s) 1816 kg PAN precursor 25.0 o C Blwr 1 tensioners 3 (g) Oxidation exhaust R1 11 (g) 3753 kg Nitrogen 1140 kg Oxygen 25.0 o C 7 (s) 300 o C Blwr 3 Tension/ speed control Mx 1 300 1000 o C 12 (g) 469.4 o C 1.0 atm exhaust 10 (g) 1300 o C Carbonization C5 1000 1300 o C C6 R3 13 (g) 750 o C C7 HX 2 C8 15 (s) 1300 o C A 14 (g) 5078 kg Nitrogen 661 kg Carbon dioxide 512 kg Oxygen 327 kg Water 12.6 kg Hydrogen cyanide 1.81 kg Carbon monoxide 1.16 kg Ammonia 0.674 kg Nitrogen dioxide 0.0101 kg Ethane 25.0 o C 2 (g) 5328 kg Nitrogen 1619 kg Oxygen 25.0 o C 8 (g) 940 kg Nitrogen 25.0 o C Blwr 2 9 (g) updated on 10/4/2010 17

21 (l) 2000 kg Water 25.0 o C P2 17 (l) 380 kg Water 20.0 kg Sulfuric acid 25.0 o C P1 18 (l) S1 HX 4 S2 19 (l) 22 (l) 25 (g) 100 o C C7 HX 5 C8 26 (l) 399 kg Water 25.0 o C A 15 (s) 1300 Ambient cooling HX 3 16 (s) 60 o C 20 (s) 24 (l) 34.7 o C 1.0 atm Mx 2 Dry 1 electricity electrolyte surface-treatment wash 23 (l) 1980 kg Water 19.9 kg Sulfuric acid 34.7 o C 27 (s) 100 o C 28 (l) 500 kg Water 10.0 kg sizing solids 25.0 o C Ambient cooling HX 8 P3 29 (l) 27a (s) 60 o C HX 6 S2 sizing 30 (l) 60 o C 31 (s) 60 o C 32 (g) 170 o C Dry 2 34 (s) 170 o C HX 7 C11 C12 Ambient cooling HX 8 33 (l) 500 kg Water 25.0 o C 35 (s) 1000 kg carbon, high strength 10.0 kg sizing solids 1.00 kg Water 0.150 kg Sulfuric acid 25.0 o C updated on 10/4/2010 18

Comments Streams Temp [C] P Phase Total Flow Water carbon, high strength Carbon dioxide Carbon monoxide Hydrogen Hydrogen cyanide Nitrogen Oxygen PAN precursor stabilized PAN Nitrogen dioxide Ammonia Ethane Ethylene Methane Sulfuric acid sizing solids Steam Water oxidized PAN R3 1 [kg/hr] R3 1 [kgmol/hr] Flow out of reactor Primary product Total conversion Per pass conversion Total yield from reactor : -4.41E- 13 93.2 1000 57.1 36.3 0.229 252 243-1762 updated on 10/4/2010 22 23.2 7.79 7.26 41.6 : 0.979 5.18 83.3 1.30 1.30 0.115 9.33 8.68-33.2 1.36 0.260 0.259 2.60 : 2702 93.2 1000 57.1 36.3 0.229 252 1183 0 0 0 0 23.2 7.79 7.26 41.6 0 0 : carbon, high strength : -3.24 NA NA NA NA NA -2.43-0 0 NA NA NA NA NA NA -0-0 : NA NA NA NA NA NA NA 100 NA NA NA NA : NA NA NA NA Error NA NA NA Error Error NA 10 1300 1.00 g 1702 93.2 57.1 36.3 0.229 252 1183 0 0 0 0 23.2 7.79 7.26 41.6 0 0 Input 11 25.0 1.00 g 4893 3753 1140 12 469 1.00 g 6594 93.2 0 57.1 36.3 0.229 252 4936 1140 0 0 0 23.2 7.79 7.26 41.6 0 0 34.5 kg Carbon monoxide is converted in rxn 1 ( 95.0 % of reactor input) 0.229 kg Hydrogen is lost in rxn 2 0.395 kg Hydrogen cyanide is lost in rxn 3 22.0 kg Ammonia is lost in rxn 4 7.26 kg Ethylene is lost in rxn 5 41.6 kg Methane is lost in rxn 6 239 kg Hydrogen cyanide is lost in rxn 7 Input to : 6594 93.2 0 57.1 36.3 0.229 252 4936 1140 0 0 0 23.2 7.79 7.26 41.6 0 0 reactor : 1.00-1.00-0.500 Reaction Coefficient 1 : 0 54.2-34.5-19.7 1 [kg/hr] : 1.23 1.23-1.23-0.616

Comments Streams Temp [C] P Phase Total Flow Water carbon, high strength Carbon dioxide Carbon monoxide Hydrogen Hydrogen cyanide Nitrogen Oxygen PAN precursor stabilized PAN Nitrogen dioxide Ammonia Ethane Ethylene Methane Sulfuric acid sizing solids Steam Water 6 [kgmol/hr] Reaction Coefficient 7 7 [kg/hr] 7 [kgmol/hr] Flow out of reactor Primary product Total conversion Per pass conversion Total yield from reactor updated on 10/4/2010 24 5.20 2.60-5.20-2.60 1.00 2.00-2.00 1.00-2.50 0 79.6 389-239 124-354 4.42 8.85-8.85 4.42-11.1 : 6594 327 0 661 1.81 0 12.6 5078 512 0 0 0.674 1.16 0.0101 0 0 0 0 : Carbon dioxide : -8.12 NA NA NA NA NA -1.42 22.8 0 NA NA NA NA NA NA -0-0 : NA NA 95.0 100 95.0 NA 55.1 NA 95.0 99.9 100 100 : NA 1115 35.0 13 750 1.00 g 6594 327 0 661 1.81 0 12.6 5078 512 0 0 0.674 1.16 0.0101 0 0 0 0 Waste 14 25.0 1.00 g -6594-327 0-661 -1.81 0-12.6-5078 -512 0 0-0.674-1.16-0.0101 0 0 0 0 0 0 15 1300 1.00 s 1000 0 1000 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 16 60.0 1.00 s 1000 0 1000 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Input 17 25.0 1.00 l 400 380 20.0 18 25.0 1.00 l 400 380 0 0 0 0 0 0 0 0 0 0 0 0 0 0 20.0 0 19 60.0 1.00 l 400 380 0 0 0 0 0 0 0 0 0 0 0 0 0 0 20.0 0 20 60.0 1.00 s 1400 380 1000 20.0 Input 21 25.0 1.00 l 2000 2000 22 25.0 1.00 l 2000 2000 wash_out 34.7 3400 2380 1000 0 0 0 0 0 0 0 0 0 0 0 0 0 20.0 0 Waste 23 34.7 1.00 l -2000-1980 0 0 0 0 0 0 0 0 0 0 0 0 0 0-19.9 0 0 0 24 34.7 1.00 l 1400 400 1000 0.150 25 100 1.00 g 399 399 Waste 26 25.0 1.00 l -399-399 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 27 100 1.00 s 1001 1.00 1000 0 0 0 0 0 0 0 0 0 0 0 0 0 0.150 0 Input 28 25.0 1.00 l 510 500 10.0

Comments Streams Temp [C] P Phase Total Flow Water carbon, high strength Carbon dioxide Carbon monoxide Hydrogen Hydrogen cyanide Nitrogen Oxygen PAN precursor stabilized PAN Nitrogen dioxide Ammonia Ethane Ethylene Methane Sulfuric acid sizing solids Steam Water 29 25.0 1.00 l 510 500 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 10.0 30 60.0 1.00 l 510 500 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 10.0 31 60.0 1.00 l 1511 501 1000 0 0 0 0 0 0 0 0 0 0 0 0 0 0.150 10.0 32 170 1.00 g 500 500 Waste 33 25.0 1.00 l -500-500 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 34 170 1.00 s 1011 1.00 1000 0 0 0 0 0 0 0 0 0 0 0 0 0 0.150 10.0 Main product 35 25.0 1.00 s -1011-1.00-0 0 0 0 0 0 0 0 0 0 0 0 0 - - 0 0 1000 0.150 10.0 Main product Overall Rxn coefficients Total yield of process (from reactant) Waste Fugitive Losses (Total) g 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Input Sum 1.75E+04 2880 0 0 0 0 0 1.00E+04 2759 1816 0 0 0 0 0 0 20.0 10.0 Fugitive Replacement 0 of Reactants Total Input (Input + Fugitive Replacement) 1.75E+04 2880 0 0 0 0 0 1.00E+04 2759 1816 0 0 0 0 0 0 20.0 10.0 Product 1011 1.00 1000 0 0 0 0 0 0 0 0 0 0 0 0 0 0.150 10.0 Sum Main 1011 1.00 1000 0 0 0 0 0 0 0 0 0 0 0 0 0 0.150 10.0 product flow Net Input (in - out, omitting fugitives) -8.53E- 06 Input C1 20.0 1.00 l 1.17E+04 1.17E+04 Cooling C2 50.0 1.00 l - 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 - out 1.17E+04 1.17E+04 Input C3 20.0 1.00 l 2.47E+04 2.47E+04 Cooling C4 50.0 1.00 l - 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 - out 2.47E+04 2.47E+04 Input C5 20.0 1.00 l 1.73E+04 2.12E+04 Cooling C6 50.0 1.00 l - 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 - out 1.73E+04 2.12E+04 Input C7 20.0 1.00 l 4.21E+04 4.21E+04 Cooling C8 50.0 1.00 l - 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 - out 4.21E+04 4.21E+04 Input C9 20.0 1.00 l 7438 7438 Cooling C10 50.0 1.00 l -7438 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0-7438 updated on 10/4/2010 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 Fug. loss kg chemical loss / hr Graph of Cumulative Chemical Losses through Manufacturing Process Cumulative Chemical Loss 1,200 1,000 800 600 400 200 0 Process Stream updated on 10/4/2010 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 kg contaminated water / hr Graph of Cumulative Contaminated Water Use / Emission through Manufacturing Process Cumulative Contaminated Water Use 3,500 3,000 2,500 2,000 1,500 1,000 500 0 Process Stream updated on 10/4/2010 28

C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 S1 S2 S3 S4 S5 S6 S7 S8 kg non-contaminated water / hr Graph of Cumulative Non-Contaminated Water Use / Emission through Manufacturing Process Cumulative Non-Contamintated Water Use 140,000 120,000 100,000 80,000 60,000 40,000 20,000 0 Process Stream updated on 10/4/2010 29

Process Diagram Label Unit Energy input [MJ / 1000 kg Product] Cumulative energy [MJ / 1000 kg Product] To [C] (Used to determine Energy Type Process diagram label Unit Energy Loss Cumulative cooling water energy Tef [C] (for recovery efficiency) Recovery Efficiency Energy Recovered Cumulative recovered [MJ / 1000 kg Product] Energy Input for each Unit Process, Cumulative Energy Requirements, Cooling Requirements (exotherms), and Assumed Heat Recovery from Hot Streams Receiving Cooling Energy Input [MJ / hr] Cooling Requirements [MJ / hr] Blw1 Blower 1 0.897 0.897 E Reactor 2-1735 -1735 300 0.600-1041 -1041 R1 Reactor 1 290 291 300 E Hx1 Heat exchanger 1-3645 -5380 300 0.600-2187 -3228 Blw2 Blower 2 0.306 291 E Reactor 4-3131 -8511 750 0.750-2348 -5576 R3 Reactor 3 6644 6935 1300 E Hx2 Heat exchanger 2-6220 -1.47E+04 750 0.750-4665 -1.02E+04 Blw3 Blower 3 0.743 6936 E Hx3 Heat exchanger 3-1984 -1.67E+04 1300 0 0-1.02E+04 P1 Pump 1 0.0231 6936 E Hx5 Heat exchanger 5-1098 -1.78E+04 100 0.250-275 -1.05E+04 Hx4 Heat exchanger 4 56.7 6992 60.0 S Hx7 Heat exchanger 7-1443 -1.93E+04 170 0.450-649 -1.12E+04 NA surface treatment 50.0 7042 0 E Hx8 Heat exchanger 8-234 -1.95E+04 170 0 0-1.12E+04 electricity P2 Pump 2 0.118 7043 E Dry1 Dryer 1 1101 8144 100 S P3 Pump 3 0.0295 8144 E Hx6 Heat exchanger 6 73.7 8218 60.0 S Dry2 Dryer 2 1402 9620 S Potential recovery -1.07E+04-1111 Net energy -1111 Potential recovery: -1.12E+04 Electricity 6986 E [MJ/hr] DowTherm 0 D [MJ/hr] Heating steam 2634 S [MJ/hr] Direct fuel use 0 F [MJ/hr] Heating natural gas 0 G [MJ/hr] Diesel process 0 Ds [MJ/hr] Undefined 0 U [MJ/hr] Heating coal 0 C [MJ/hr] Energy input 9620 [MJ/hr] updated on 10/4/2010 30

requirement Cooling water -1.95E+04 [MJ/hr] Cooling refrigeration [MJ/hr] Potential heat -1.07E+04 [MJ/hr] recovery Net energy -1111 [MJ/hr] updated on 10/4/2010 31

Start Blower 1 Reactor 1 Blower 2 Reactor 3 Blower 3 Pump 1 Heat exchanger 4 surface treatment electricity Pump 2 Dryer 1 Pump 3 Heat exchanger 6 Dryer 2 Potential recovery MJ / hr Graph of Cumulative Energy Requirements Cumulative Energy Input 12,000 10,000 8,000 6,000 4,000 2,000 0-2,000 Process Unit updated on 10/4/2010 32