Carbon Materials for Advanced Technologies Episode 6 ppt

180 their study. The solid carbon thermal conductivity must be scaled to account for density, and the fact that not all of the solid carbon contributes to the solid thennal conductivity. The temperature dependence of the thml conductivity of graphite is typically modeled as (3) =A T-e 1 where A, and E are fitting parameters, with e usually being set to 1 for perfect graphite. For polygranular graphite, which contains a significant numbers of point defects and other phonon scattering sites, E is between 0.5 and 1. Less graphitic carbons, and graphites with lower graphitization temperatures that have a significant population of extended crystal lattice defects, have E values of less than 0.5. For CBCF, Dinwiddie et al. decided to modify Eq. (3) for the solid: where X and Z are functions of the heat treatment temperature. A, and E’ were determined by fitting the thermal conductivity data from “as fabricated” CBCF specimens (Fig. 6) where X= 1, Z = 0, and A, and e‘ were found to be 0.02328 and 0.2380, respectively. Combining Eq. (1) and (4) gives the functional form of the equation for the thermal conductivity of CBCF as h. = XAIT(e’Q + 2.160 x lO-l1T3 Equation (5) was fitted to experimental data for the thermal conductivity of CBCF heat treated at various temperatures for 10, 15, and 20 seconds, and a linear relationship was determined for Z and the heat temperature, T,, which is given by E¶. (6). Z = 1.276 - 0.0005659 x T, Z varies from - -0.23 to - - 0.64 over the heat treatment temperature range 2673 to 3273 K, and the term (&‘+a in Eq. (5) therefore varies from -0 to 0.4 over the same temperature range. Equation (6) was then substituted back into eq. (5) which was used to refit the data to determine the relationship between X and the heat treatment conditions (time and temperature, Eqs. 7 and 8). The empirical parameter Xin Eq. (5) was found to be given by X = R * exp(Tm x 0.004963) 181 Q where R is a time dependent variable given by R = 1.261 x + 3.617 x x t (8) where t 1s the tune at the heat treatment temperature. By definition X is unity when tis zero (ie., in the “as fabricated” condition), and from Eq. (8) we see that R = 1.261 x lO-’when t is zero. To satisfy this condition for R, t must take the value of 4.3 seconds. It was necessary to subtract 4.3 seconds from the time m the furnace hot zone in order to bmg the R parameter value in line with the “as fabricated” value. Physically, the 4.3-second correchon can be considered to be the time taken for the CBCF to equilibrate at the furnace temperature. Figures 7 and 8 show thermal conducbvity data for CBCF after exposure to temperatures of 2673, 2873, 3073, and 3273 K, for 5.7 and 15 7 seconds, respectively. The symbols in the Figs. 7 and 8 represent measured thermal conductivity values, and the solid lines are the predicted behavior from Eqs. (5) through (8) The model clearly accounts for the effects of measurement temperature, exposure me, and exposure temperature The fit to the data is good (typically within 10%). However, the fit to the “as fabricated” CBCF data (Fig 4) was less good (- 20%), although the scatter in the data was larger because of the much lower heat treatment temperature (1 873 K) in that case. 4 Damage Tolerant Light Absorbing Materials The optical properhes of low density, carbon fiber-carbon bmder composites have recently been disclosed by Lauf, et a1 [14,17]. CBCF samples were fabncated accordmg to the method described m Section 2 of this chapter, and were tested to determine their optical properties (light absorpbon and spectral reflectance). The optxal scatter was measured at a wavelength of 10.6 pm for scattemg angles from 0.1 to 100 degree from specular. The absorbance, measured as the bidirecbonal reflected distribution function (BRDF), of abraded CBCF is shown as a fhcbon of scattering angle in Fig. 9 for the parallel and perpendicular to moldmg hections. Significantly, the material is essenhally isotropic wth respect to its BRDF. Also included in Fig. 9 are data for a commercial light absorbmg product, namely Martin Black 54, an alummum with a black-anodized surface The CBCF and Martin Black 54 &play similar light absorption properhes (both appear to be totally Lambertian) showing no specular scattering. The spectral reflectance of 182 CBCF for infrared wavelength from 2 to 55 pm is compared with that of etched beryllium in Fig. 10. The data show that the CBCF IS uniformly light absorbing up to at least 50 pm, in marked contrast to the etched beryllium light absorber whch effectively absorbs wavelengths only up to about 20 pm. .I .3 1 3 10 30 100 SCA~R!N@ ANGLE (degrees from specular) kio.qm (PL=300 Fig. 9. Light absorption behawor as a funchon of scattering angle for CBCF III the 11 and I to molding direction, and Martin Black 54 [ 161. Light absorbers must be materials that are very absorbent (or “black”) over the widest possible range of wavelengths, ideally including the infrared spectrum, so as to be more effective in sensitive, precision optical system. Typically, hght absorbers are made from light metals, e.g., beryllium or aluminum, and derive their light absorbing properties from a mcroscopically textured surface coating developed on the metal via a chemical etching or anodizmg process. Martin Black 54 is one such material. Martin Black 54 is one of the better light absorbers, exhibiting excellent absorpbon behavior and very low scatter throughout a broad range of optical and infrared wavelengths. However, the surface of Martm Black 54, and other anodized light absorbers, is very fragile. Once the surface is damaged, the hght absorbing properbes quickly dirrrrmsh and the materials loses its effectiveness as a light absorber. Etched beryllium light absorbers are somewhat more robust than Martm Black 54 but, as shown rn Fig. 10, are meffecbve above certam wavelengths. Moreover, both beryllium and aluminum are sensitive to environmental degradation and may degrade thermally due to their low melting temperatures. 183 75 50 25 0 2 14 26 38 50 62 WAVELENGTH (pm) Fig. 10. The spectral reflectance of CBCF and etched beryllium over the wavelength range 2 to 55 pm [16]. In contrast to the above coated metal absorbers, CBCF is rugged and durable. Its optical properties have been shown to be independent of surface condition because they derive from the properties of the bulk material and not the surface condition. CBCF can be readily machined and is easily attached by adhesive bonding or brazing to metallic or graphite substrates. These attributes make CBCF ideally suited for applications such as broadband optical absorbers, baffles, beam stops and other structures to protect detectors from stray light at high power and long wavelengths. Moreover, the combination of the above properties and the uniform emissivity of CBCF (0.62 - 0.63 at ambient temperature), whch is independent of surface roughness [ 171, makes it a suitable measurement or calibration standards for spectrophotometers, scatterometers or other optical instruments [ 181. 5 Carbon Fiber Composite Molecular Sieves 5.1 Applications A recently developed adsorbent version of ORNL’s porous carbon fiber-carbon binder composite is named carbon fiber composite molecular sieve (CFCMS). The CFCMS monoliths were the product of a collaborative research program between ORNL and the University of Kentucky, Center for Applied Energy Research (UKCAER) [19-211. The monoliths are manufactured in the manner described in Section 2 from P200 isotropic pitch derived fibers. While development of these materials is in its early stages, a number of potential applications can be identified. 184 It is anticipated that these materials will only find ublity in applicabon that can support the relatively high cost of their manufacture compared to commodity granular acbvated carbon (GAC). However, the monoliths may also fiid applications in situations were their novel properties make them uniquely suited. Potenbal areas of application include gas separation and cleanup, especially m the field of air purity for buildings or vehicles. In thw latter area, collective protection systems for military equipment [against nuclear, biological and chemical (NBC) threats] would appear to be a promising application. At UKCAER, researchers have shown the monoliths be extremely effective at removing a common herbicide (sodium pentachlorophenolate, or PCP) from water [22], offermg a potenbal application in ground water cleanup systems. Another possible application of our monohths is adsorption gas storage, where the potential for high CFCMS bulk density, combmed with high micropore volume and high deliverable gas capacity, makes them attractive. The novel electrical desorption capability of the material (Sect. 5.3), combined with the umque pore structure of the monolith, make the matenal particularly suited to utility as a guard bed for adsorbed natural gas (NG) storage tanks (e.g., on a NG powered vehicle), or for a NG fueled solid oxide fuel cell. The properties, pore smcture, and performance of the monoliths are described below and their suitabdity to specific applications is discussed. 5.2 Pore Structure 5.2.1 The structure of unactivated CFCMS monoliths The macrostructure of a CFCMS monolith is shown in Fig. 1 1. The isotropic pitch- denved carbon fibers have a smooth surface and a circular cross-section, whch 1s in marked contrast to the rayon-derived fibers used in the manufacture of CBCF. The fibers are bonded at their contact points by the carbonized phenolic resin, thus forming a continuous three dimensional carbon network. The carbon fibers are 12- 14 pm in diameter and, in the monolith shown in Fig. 11, have an approxlmately normal length distribution with a mean of -450 pm. The fiber length distribution mode is -400 pm and the fiber length varies widely from 100 to 1000 pm [23]. The voids between the fibers are typically >30 pm in sue and the resultant open structure allows the free flow of fluids through the material. Mercury porosimetry data taken on unactivated CFCMS material are shown in Fig. 12, and mdcate the macropore (inter-fiber voids) size range to be approxlmately 10 to 100 pm, in good agreement wth the macropore sizes indicated in the SEM mcrograph (Fig. 11). The absence of mesoporosity (2-50 m) can also be noted in Fig. 12. In the unactrvated condtion, CFCMS contains a small micropore volume that presumably develops during the carbonization stage. Typical DR micropore volumes and BET surface areas for the unactivated monoliths are 0.01 to 0.04 cm3/g and 70 to 100 m2/g, respectively. 185 Fig.11. SEM micrograph showing the macrostructure of CFCMS (scale bar represents 1 OOpm). =j 3.5 - - Y E 3.0 c u) .O 2.5 2 - E 2.0 100 10 1 0.1 0.01 Diameter (pm) Fig. 12. Mercury intrusion pore size distribution for an unactivated sample of CFCMS monolith. 186 5.2.2 Microporous CFCMS monoliths The development of microporosity during steam activation was examined by Burchell et al. [23] in their studies of CFCMS monoliths. A series of CFCMS cylinders, 2.5 cm in diameter and 7.5 cm in length, were machined from a 5- cm thick plate of CFCMS manufactured from P200 fibers. The axis of the cylinders was machined perpendicular to the molding direction ([[to the fibers). The cylinders were activated to burn-offs ranging from 9 to 36 % and the BET surface area and micropore size and volume determined from the N, adsorption isotherms measured at 77 K. Samples were taken from the top and bottom of each cylinder for pore structure characterization. Full accounts of hs study can be found elsewhere [23-251. However, the results are summarized in Table 1, where mean values for the surface area and pore parameters are given for each cylinder. The pore size and volume increased wth mcreasing burn-off, as did the BET surface area, in agreement with previous pore structure development studies conducted on this material [26]. The variation of pore volume and surface area was noted to be particulary large m the data for high burn-off samples (C25%), which was attributed to uneven activahon in the monoliths [23]. For example, the BET surface area was noted to vary by almost a factor of three over the 7.5 cm specimen length. In an attempt to achieve uniform activation of larger (10 cm in diameter and 25 cm in length) monoliths, an oxygen chemisorptiodactivation procedure was adopted (Section 2). The monoliths were subjected to two cycles of 0, chemisorptiodactivation and attamed total burn-offs of approximately 8.5-13.4 %. One of the monoliths was secboned and sampled to determine the spatial vanation of the BET surface area, and micropore volume and slze. Samples were taken at several radial locations across slices cut periodically along the length of the monolith. The spatial variation of the DR micropore volume is shown in Fig. 13. Table 1. Micropore structure development during steam actwation for CFCMS monoliths manufactured from P-200 carbon fibers. ~~~ Pore volume DA Pore Burn-off BET Area [t-method] radius Specimen ("/.I (m%) (cm3/g) (nm) 21-11 9 5 12 0.21 0.68 2 1 -2B 18 1152 0.40 0.71 2 1 -2c 27 1962 0.65 0.75 2 1 -2D 36 1367 0.41 0.79 187 fl 0.15 - V Fig.13. Vanation of DR micropore volume in a large CFCMS monolith wth 10 4 wt% bum- off [27]. The micropore volume varied from -0.15 to -0.35 cm3/g. No clear trend was observed with respect to the spatial variation. Data for the BET surface area are shown in Fig. 14. The surface area varied from -300 to -900 m2/g, again with no clear dependence upon spabal location withm the monohth. The surface area and pore volume varied by a factor -3 withm the monolith, which had a volume of - 1900 cm3. In contrast, the steam activated monolith exhibited slmilar mcropore structure variability, but in a sample with less than one fiftieth of the volume. Pore size, pore volume and surface area data are given in Table 2 for four large monoliths activated via 0, chemisorption. The data in Table 2 are mean values from samples cored from each end of the monolith. A comparison of the data m Table 1 and 2 indicates that at burn-offs -10% comparable pore volumes and surface areas are developed for both steam activabon and 0, chemisorptiodacbvation, although the process time is substanbally longer m the latter case. Table 2. Micropore structure data for large CFCMS monoliths activated via the 0, chemsorptiodactivation route [7,27] BET surface DR micropore DA Monolith Bum-off area volume mcropore ID (“/.I (m’k) (cm’ k) diameter (nm) ____ ~~ ~ SMW-1 10 496 0 22 159 SMW-3 9.4 528 0 20 1 46 SMW-4 8.5 714 0 27 1.55 SMW-8 13.4 574 0.22 1.51 188 Fig. 14. BET surface area as a function of position in a large CFCMS monolith wth 10.4 wt% bum-off [27]. 5.2.3 Mesoporous CFCMS monoliths The pore structure of monoliths made from F3-0 PAN-derived carbon fibers has been examined and found to be highly mesoporous [18,29]. The cumulative mesopore surface area, as a hction of pore diameter, is shown m Fig. 15 for the PAN fibers, and monoliths in the dried (50°C) and carbonized (650°C) conditions. The surface area is clearly associated with pores of slze <50 nm, ie., the mesopores. The carbonized monoliths exhibit surface areas >500 m2/g and mesopore volumes >1 cm’/g. In contrast, the “as received” PAN fibers exhibited a mesopore volume of only 0.28 cm3/g and surface area <200 m2/g. The mcrease in mesopore volume and surface in the monoliths was attributed to the openmg of the mer pore structure of the fiber through gasificahon by reactive species such as 0,, CO,, H,O, and CO, which are thought to be adsorbed during monolith production and subsequently desorbed during carbonizahon [28,29]. The PAN fiber monoliths were subjected to steam activahon at 650°C or 850°C to burn-offs up to -22%. The mesopore volume was observed to reduce siglllficantly with burn-off (Fig.l6), reachmg a lower llmit of -0.3 cm3/g above - 10% burn-off A similar trend was observed for the mesopore surface area. Conversely, the mean mesopore diameter mcreased from -7 to 8.7 nm over the same bum-off range. 189 700 600 - ," 500 - 400 - - E a 5 a, 200 - a, X 300 - u) > 100 - * 0- -100 1 SEM examination of the steam activated PAN fiber monoliths showed the fiber diameter to be significantly reduced during the activation process, suggesting the fibers are consumed radially by a gasification process of the external surface [28]. Total Mesopore Suface Area -PAN Fibers Carborized 650°C 572 -? L h I111.11.1 3 .ll*,lll I 111. 9 1 11.11 I1111 1 .1>.111 1.2 1 .o '" 7 0.8 - - 5 0.6 - 9 % -g 02- - E I aJ $ 0.4 - m 0.0 Fig.15. Mesopore surface area as a function of pore diameter obtained from mercury intrusion data for PAN derived carbon fiber porous monoliths [28]. 0 650°C 850'C 0 0 0 0 * e+ e I I I I Fig.16. Mesopore volume as a function of burn-off for PAN derived carbon fiber porous monoliths [28]. [...]... New f York, 1980, pp 1032 1042 Wei, G.C and Robbins JM., Carbon- bonded carbon fiber insulation for radioisotope space power systems, Ceramics Bulletin, 1985 ,64 (5) ,69 1 69 9 202 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 Weaver, C E and Chilcoat, B.R., Carbon- bonded carbon fiber for space applicahons Paper presented at the 1994 American Carbon Society Workshop, Oak Ridge, TN., May 15-18, 1994... on Carbon, American Carbon Society, 1979, pp 247 248 Donnett, JB and Bansal, R.C., Carbon Fibers, 2nd edition, Marcel Dekker, Inc., New York, 1990 Jagtoyen, M and Derbyshire, F., Carbon fiber composite molecular sieves for gas separation In Proc Tenth Annual ConJ on Fossil Energy Materials, CONF 960 5 167 ,0RNWFMP- 96/ 1 Oak Ridge National Laboratory, 19 96, pp 291 300 Nandi, S.P and Walker, P.L Jr., Carbon. .. Matheny, M and Burchell, T., Carbon fiber composite molecular sieves for gas separation In “NewHorizons for Materials Advances in Science and Technology (edited by P Vincemni), Techna Srl, Faenza, Italy 1995, Vol 4, pp 41 1 417 Jagtoyen, M., Lafferty, C., Kimber, G, and Derbyshire, F., Novel activated carbon materials for water treatment In Proc The CARBON ‘ 96 Conf, 19 96, pp 328 329 Burchell, T D.,... in this context is discussed subsequently 194 Table 5 Micropore characterization data for hybrid monoliths at two densities Specmen ID Pre-activation density Wm’) % DKDX fibers Bum-off (“w BET area (mZ49 DR pore vol (cm’k) KOA 0 .67 0 5.5 429 0. 16 K1A 0 .62 5 72 4 06 0 16 K3A 0 .69 18 43 307 0 12 KOB 0.21 0 56 445 0 16 KlB 0.22 5 94 540 0.21 K3B 0.25 18 5.3 429 0.17 5 4 Gas adsorption and separation The... process and material for the separation of carbon dioxide and hydrogen sulfide gas mixtures Carbon 1997, 35(9), 1279 1294 Burchell, T.D., Judkins, R.R, Rogers, M.R and Williams, A.M A novel approach to the removal of COP In Proc Tenth Annual Con$ On Fossil Energy Materials, ORNL/FMP- 96/ 1, COW- 960 5 167 , Oak Ridge National Lab, U.S A , 19 96, pp 135 148 Burchell, T.D and Judkins, R.R A novel carbon fiber based... Energy Materials, ORNLRMP-95/1, CONF-9505204, Oak Ridge Nahonal Lab, U S A , 1995, pp 447 4 56 Klett, J.W and Burchell T.D., Carbon fiber carbon composites for catalyst supports In Proc 22"dBiennial Conf On carbon, Pub Amencan carbon Society, 1995, pp 124 125 M Kuro-Oka, T Suzuki, T Nitta and T Katayama J Chem Eng ofJapan, 1984, 17 (6) , 588 Smgh, P., Ruka, R.J., and George, R.A Direct utilization of hydrocarbon... York, 1989, pp 1553 1 563 205 CHAPTER 7 Coa1-Derived Carbons PETER G STANSBERRY, JOHN W ZONDLO, ALFRED H STILLER Department of Chemical Engineering West Virginia University Morguntown, WV 26. 5 06- 6102 1 Review of Coal-Derived Carbons 1 I Introduction Complex aromatic raw materials such as petroleum resids, decant oils, coal, and coal tars have been employed for many years by the carbon industry and contlnue... Burchell, T D., Carbon fiber composite molecular sieves In Proc Eighth Annual Conference on Fossil Energy Materials, ORNL/FMP-94/1, CONF9405143, Oak Ridge National Lab, U.S.A., 1994, pp 63 70 Burchell, T.D., Weaver, C E., Derbyshire, F., Fei, Y.Q and Jagtoyen M., Carbon fiber composite molecular sieves:synthesisand charactenzahon In Proc Carbon ‘94,Granada, Spain, Spanish Carbon Group, 1994, pp 65 0 65 1 Derbyshire,F.,... removal using a carbon fiber composite molecular sieve, Energy Convers Mgmt, 19 96, 37 (6- 8), 947 954 Burchell, T D and Rogers, M.R., Carbon fiber composite molecular sieves, In Proc Eleventh Annual Con$ On Fossil Energy Materials, ORNL/FMP-97/1, CONF-9705115, Oak Ridge National Lab, U.S.A., 1997, pp 109 1 16 Burchell, T.D, Klett J.W ,and Weaver, C.E A novel carbon fiber based porous 203 29 30 3 1 carbon monolith,... the uniformity, and key properties of the material, and at containing, or reducing, the cost of our porous carbon fiber -carbon binder composites 7 Acknowledgments Research sponsored by the U S Department of Energy under contract DE-ACOS 960 R22 464 with Lockheed Martin Energy Research Corporation at Oak Ridge National Laboratory 8 References 1 2 3 4 5 6 7 8 9 10 Ardery, Z.L and Reynolds, C.D., Carbon . determined for Z and the heat temperature, T,, which is given by E¶. (6) . Z = 1.2 76 - 0.000 565 9 x T, Z varies from - -0.23 to - - 0 .64 over the heat treatment temperature range 267 3. data for PAN derived carbon fiber porous monoliths [28]. 0 65 0°C 850'C 0 0 0 0 * e+ e I I I I Fig. 16. Mesopore volume as a function of burn-off for PAN derived carbon. (“w (mZ 49 vol. (cm’k) KOA 0 .67 0 5.5 429 0. 16 K1A 0 .62 5 72 4 06 0 16 K3A 0 .69 18 43 307 0 12 KOB 0.21 0 56 445 0 16 KlB 0.22 5 94 540 0.21 K3B 0.25 18 5.3 429 0.17 5