Carbon Materials for Advanced Technologies Episode 6 ppt

= XAITe’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 linea

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 t h m l conductivity of graphite

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) Q

where R is a time dependent variable given by

where t 1s the tune at the heat treatment temperature By definition X is unity when

t i s 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 b m g 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 m e , 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 f h c b o n

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

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

S C A ~ R ! N @ ANGLE (degrees from specular)

k i o q m

(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

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

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 C F C M S monoliths

The macrostructure of a C F C M S 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 w t h 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

Fig.11 SEM micrograph showing the macrostructure of CFCMS (scale bar represents

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 h s 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

- 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

Fig 14 BET surface area as a function of position in a large CFCMS monolith w t h 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 h c t i o n 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 m e r 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

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

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

5.3 Physical Properties

Typical physical properties of our CFCMS monoliths are given in Table 3 The exact value of a parbcular property is dependent upon the fabrication route, composition, density, etc Consequently, property ranges are given m Table 3

rather than absolute values

Table 3 Physical properhes of CFCMS monoliths developed by ORNL/UKCAER

* hot-pressed density range

The effect of fiberbinder ratio on the density and strength of the isotropic pitch derived fiber monoliths was examined [23] in a study 111 which the raho of P200 fibers was increased by factors of 2, 3, or 4, from the standard fiberhmder ratio The density was seen to decrease from -0.38 g/cm’ for the standard formulation

to -0.36 gkm3 for the 4X fomulahoa A slight reductmn in compressive strength,

from - 1.9 to - 1.7 MPa, was observed to accompany the density reduchon, although the scatter in the data made it impossible to develop a density-strength conelabon D m g activatlon the carbon is selectively gasified, resultmg 111 a mass loss m the monolith The compressive strength (uc) is degraded by this mass loss and follows the relahonshp [ 19,231:

(9)

oc = 1.843exp(-O.O1323x)

where x is the fracbonal weight loss or burn-off

The electncal behavior of CFCMS IS shown ~tl Fig 17(a) The current-voltage relationship is linear and the electrical resistwity of the monolith in Fig 17(a) (2.5

cm in diameter and 7.5 cm in length) was 130 d - c m [23,24] The resistivlty of the monolith is considerably greater than that of the carbon fibers, whch according

to the manufacturer’s product literature 1s 4-6 mS2.cm The poorer electncal conductivity of the monolith can be attributed to the large electncal resistance

associated with the fiberibmder interface A consequence of the passage of an electric current through the monolith is resistwe (ohrmc) heatmg Figure 17(b)

shows the temperature of a monolith as a function of the electrical power input (product of the applied voltage and induced current) At relatwely low power inputs, the monolith readily heats to 50-100°C, and temperatures >300°C are rapidly attained at a power input of -45 W

Fig 17 The voltage-current relationship (a) and resistive heating curve (b) for a CFCMS

monolith (sample 21-2B, 18% burn-off, 2.5-cm diameter x 7 5-cm length) [23]

Data for the thermal conductivity of adsorbent carbons are somewhat limted [23] Typically, a bed of granular carbon at a packed density of -0.5 g/cm3 has a thermal conductwity of 0.14-0.19 WImK, while the denved value for the carbon adsorbent

is normally between 0.6 and 1.3 WlmK CFCMS monoliths typically have

comparable thermal conductivities to a packed bed, but at substantially lower density (Table 3) The greater specific thermal conductivity of the monoliths can

be attributed to the substanhally higher thermal conductwity of the carbon fibers (2-5 W/mK), whch results from the higher density of fiber compared to GAC (1.6 g/cm3 cf 0.6 g/cm3), and the reduced contact resistance between the fibers in the case of the bonded fiber monoliths For many applicabons increased thermal conductmty is an extremely desirable attriiute for a bed of adsorbent carbon The flexible process by which CFCMS is manufactured allows the blending of high conductivity mesophase pitch-derived carbon fibers into the material Moreover, hot pressmg the monolith after drylns allows the density and thermal conductivity

to be increased substantially To assess the extent to which the thermal conductwity could be enhanced by blendmg m mesophase pitch-denved carbon fibers, andor by increasmg the bulk density, a series of expermental hybrid monoliths were fabricated Table 4 reports the compositlon, density, and room

temperature thermal conductivities of the monoliths

Table 4 Room temperature thermal conductivity of hybrid monoliths at normal and hlgh density

Thermal conductivity Specimen wt % of DKDX Density at 25°C (W/mK)

as the fraction of DKDX fiber m the hybrid monolith increases At a loadmg of 18% DKDX fibers, the thermal conductivity (11) is increased to 0.19 W/mK at a density of 0.26 g/cm3 and 0.93 W/mK at a density of 0.63 g/cm3 T h s latter value represents a six-fold increase over the thermal conducbvity of the standard CFCMS monoliths and a four- to suc-fold mcrease over the thermal conduchvity of a packed bed of GAC The temperature dependence of the thermal conductivity (11) of the hybrid monoliths is shown in Fig 18 The thermal conducbvity increases with

temperature over the range 30-500°C due to the increasing contnbubon of radation conduction m the pores (see the discussion in Secbon 3 of this chapter)

An increased thermal conductivity in a carbon bed will reduce temperature gradients, q r o v e efficiency and, for a storage carbon, will increase the delivered capacity of gas If, however, the mcreased thermal conductivity is accompanied

by a large reduction in bed adsorpbon capacity, the potential performance gain may

be totally offset by the capacity loss penalty To assess the extent, if any, of this potential penalty the hybrid monoliths were activated via the 0, chemisorptiodactivabon process and their micropore structure examined Table

5 reports micropore characterizabon data for the hybrid monoliths (standard and

required to increase the micropore volume and storage capacity of the monoliths Some of our preliminary work in this context is discussed subsequently

Table 5 Micropore characterization data for hybrid monoliths at two densities

5 4 Gas adsorption and separation

The gas adsorption behavior of our monoliths has been studied as part of the U.S

Department of Energy’s ongoing Fossil Energy Advanced Research Program The equihbrium adsorption of CO, and CH, was found to be strongly temperature dependent, and the uptake of CO, was greater than the uptake of CH, for a given

specimen [23] For example, volumetric measurements at 30°C and one atmosphere, on CFCMS with moderate burn-off, showed that approximately 50

cm3/g of CO, were adsorbed, whereas only approxmately 27 cm3/g of CH, were

adsorbed High pressure [0.5-59 bar (8-850 psi)] CO, and CH, isotherms are shown in Fig 19 for monoliths 21-1 1 and 21-2B, which had 9 and 18% burn-off, respectively The measured volumetric and gravmetric (Fig 19) adsorption capacities at one atmosphere for both CH, and CO, are in good agreement for the CFCMS specmens At one atmosphere, approxmately 100 mg of CO, per g of CFCMS and approximately 19 mg of CH, per g of CFCMS were adsorbed The quantihes of gas adsorbed rose to >490 mg/g (CO, on specimen 21-2B) and >67 mg/g (CH, on specimen 21-2B) Moreover, the CO, isotherms are still mcreasmg with pressure whereas the CH4 isotherms have flattened (i.e., the CFCMS has become saturated with CH,) The data in Fig 19 clearly show that CFCMS exhibits selective adsorption of CO, over CH,

breakthrough apparatusC23-2.51 A typical breakthrough plot for a CH,/CO,

murture is shown m Fig 20 The specimen is heated electslcally and any entrained air is initially dnven out with a He purge The mput gas is then switched to a 2: 1 mixture of CH,/CO, at a flow rate of 0.33 slpm The outlet stream He concentration decreases and the CH, concentration increases rapidly (i.e., CH, breaks through) Adsorption of CO, occurs and, therefore, the CO, concentrahon remains constant at a low level for apprownately six minutes before the CO, concentration begins to increase, i.e., CO, -breakthrough occurs Table 6 reports

data from a prelminary study of CO, separation CO, capacibes are reported as determined from pure CO, and CO,/CH, mixtures on each speclmen examined The reported CO, capacibes are the means of several repeats of the breakthrough expenments, and the BET surface and other microporosity charactenzation data are addiQonally given m Table 1 Two of the CFCMS samples (lowest bum-off) had

CO, adsorption capacities of almost one liter on 0.037 bters of adsorbent, and only

a small capacity reduction was observed in the COJCH, gas murture The CO, adsorption capacity decreases with mcreasing burn-off, in agreement with the isotherm data

Table 6 CO, seoaration data from our CO, and COJCH, breakthrough exoeriments

Specimen Bum-off BET Surface CO, Capacity (Liters)

to increase sharply after breakthrough is completed The concentration mcrease

coincides with the applicatlon of a d.c electrical voltage (4-5 amps at 1 volt) and the He purge gas H2S desorption occurs over a relatively short tune (1 8 minutes) The H,S adsorption capacity (at atmospheric pressure) for sample 21-1B, 18%

burn-off, was 0.43 liters (Fig 21)

Flow rate = 0.44 slpm Gas composition: HPS 5.4%

to affect a rapid desorption of adsorbed gases in our breakthrough apparatus The

benefit of this technique is shown m Fig 22, which shows the CO, and CH, gas concentrations in the outlet gas stream of our breakthrough apparatus [23-251

Three adsorptionidesorption cycles are shown ~ f l Fig 22 In the f i s t and second

cycles (A and B in Fig 2 2 ) desorption is caused by the combined effect of an

applied voltage (1 volt) and a He purge gas In the third cycle (C in Fig 22)

desorphon is caused only by the He purge gas A comparison of cycles B and C mdicates that the applied voltage reduces the desorptlon tune to less than one third

of that for the He purge gas alone (cycle C) Clearly, the desorphon of adsorbed

C 0 2 can be rapidly induced by the apphcabon of a d.c electncal potenfial

Fig 22 CO,/CH, breakthrough plots for CFCMS sample 2 1 - 1 F (1 0% bum-off) showing

He purge @? 0.06 slpm, and C 0 volt, He purge 0.06 slpm

Increased adsorbent (CFCMS) temperature results m desorphon of the adsorbate However, desorption occurs m e d i a t e l y when the voltage is applied to the CFCMS, whereas the bulk temperature increase lags the apphcahon of the voltage

by a finite time, typically several minutes [23] Evidently, the resistance heatmg effect is acting directly at the adsorption sites (fiber mcroporosity) resulting m a rapid desorption of the adsorbate The heat of adsorption of CO, on activated carbon fiber is 30 kJ/mol [30] A simple calculation for the a typical ESA

breakthrough experiment, where approximately 1 litre of CO, was adsorbed, mdicates that at a power level of 5 Watts, approxunately 270 seconds would be

required to mput the energy (1350 J) required to desorb the C02 adsorbed on the

CFCMS Implicit in this calculation is the assumption that all of the electrical energy is converted to thermal energy and transferred to the adsorbed CO, Whlle this analysis is very smplistic, it does explam the observation of a t m e lag between the inihation of electrical current flow and the CFCMS temperature rise during the electrical desorpfion of adsorbed gases Actual measured desorphon tunes are of the order of 6-13 minutes, depending on the purge gas flow rate (Fig 22) Therefore, other factors must lnfluence heat flow to the adsorpfion sites in the

carbon fibers Several explanations have been postulated, the most plausible of which is based on the compensating effect of the heat of adsorption and the temperature dependence of electrical resistivity in carbon [23]

The ability of CFCMS to selectwely adsorb a gas from a gas rmxture, combined with the electncally enhanced desorpbon of the adsorbed species, allows for a gas separabon system where the separation is effected by electrical swmg (ES) rather than the more conventional pressure or temperature swings Several applicahons

of CFCMS/ES can be considered For example: (i) the cleanup of sub-quality natural gas; (ii) the separation of hydrogen from coke oven battery reformer waste gas streams; (iii) the separation of landfill gases; and (iv) a guard bed for the removal of higher hydrocarbons, or sulfur bearlng odorants, from CH, fuels in adsorption storage fuel tanks or solid oxide fuel cells Moreover, the novel combmation of properties make CFCMS attractive for adsorption gas storage systems where the delivery of adsorbed gasses can be hindered by excessive

temperature drop in the carbon adsorbent due to the large heat of adsorption A

variant of the CFCMS monolith with appropriately developed microporosity, and

a b u k density - 1 O g/cm3, would be expected to posses a storage volume equal to

or greater than currently available CH, storage carbons Fmally, a mesoporous variant of CFCMS might offer advantages as a catalyst support for reforming operations because heating of the support could be effected directly by the passage

of an electric current, negatmg the need to preheat the reactant gasses

5.5 Near term applications

Two particular applications of CFCMS monoliths can be considered near term The first, fighting vehicle air clean-up (with respect to NBC contarmnants), would appear to be an eminently suitable field for our adsorbent fiber-based monohths Several attributes of the monoliths should be considered in this context: (1) the monohths are rugged and wll not suffer attntion under the harsh terrain condihons encountered by fighting vehicles; (ii) the micro/meso pore structure can be controlled by appropnate selecbon of the fiber type and processmg/activation route; (iii) ESA would appear to offer a rapid and low energy desorptiodregeneration method, compared to pressure swing or thermal swing regeneration for the adsorbent bed; and (iv) the defense market could stand the higher cost of the monoliths compared to granular carbons The second near term application of our adsorbent monoliths is in a guard bed for a solid oxide fuel cell (SOFC) [7,27] Westinghouse solid oxide fuel cells utilize CH, and air as fuels [31] Operatmg experience with the cells has demonstrated an efficiency degradation associated with the interaction of the sulfur bearing odorants in the natural gas and the ceramic materials used m the construction of the cell Thls has necessitated the use of a large GAC guard bed, which must be replaced when saturated A compact, easily regenerated guard bed has obvious advantages over