Adsorption of So2 on Bituminous Coal Char

7/29/2019 Adsorption of So2 on Bituminous Coal Char

1/5

I

\

I

ADSORPTION OF SO2ON BITUMINOUS COAL CHAR AND ACTIVATED CARBONFIBER PREPARED FROM PHENOL FORMALD EHYDEJosep h A. DeBarrl.', Anthony A. Lizzio' and M ichael A. Daley''Illinois State Geological Survey, 615 E. Peabody D r., Cham paign, IL 61820'University of Illinois, Environmental Engine ering and Scie nce Program, Urban a, IL 61 801'University of Illinois, Materials Science and Eng ineering Dep artmen t, Urba na, IL 61801

Keywords: A ctivated carbon fiber, coal char, SO, adsorptionINTRODUCTIONCarbon-based materials are used comm ercially to rem ove SO, from coal combustion flue gases.Historically, these materials have consisted of granular activated carbons prepared from lignite orbituminous coals 11-31, Recent studies have reported that activated carbon fibers (ACFs) may havepoten tial in this application d ue to their relatively high SO, adsorp tion capacity 14-61, In this paper, acomparison of SO, adsorption for both coal-based carbons and ACFs is presented, as well as ideas oncarbon properties that may influence SO , adsorption.EXPERIMENTALSamole PreoarationThe chars used in this work were prepared from an Illinois No. 2 hvCb coal, sample IBC-102 of the IllinoisBasinCoal SampleProgram 71. A physically cleaned 48x100 mesh sample having 3.6% mineral contentwas prepared from the parent coal and used throughout as the feedsto ck Tor activated char. Chars wereprepared at 900C for 0.5 h in a 5 cm ID batch fluidized-bed reacto r (FER). In each run, 200 g IBC-102coal was fluidized in flowing N, 6 Wmin) and heated to the desired pyrolysis temp erature. A multistepheating p rocedure was used to minimizeagglomeration of coal pa rticle s in the FBR. Steam activation wasdone to develop microporosity and increase surface a rea further. Typically, 50 g of char was placed in theFBR and heated to 860 C in flowing N? Th e N, flow was replaced by 50% H,0/50% N, (6 Wmin) for0 75 h to ach ieve 30% carbo n conversio n. In some cases, the H,O-activated char was treated with nitricacid (HNO,). Typically, 10 g of the char was added to 0.2 L I O M HNO, solution, and refluxed at R O Tfor 1 h. The H N0,-treated carbon was washed with distilled H,O to remove excess acid and vacuum driedovernight at 25 T . In some cam, he HN0,-treated char was heated in N2 o 525,725, or 925T and heldfor I h to remo ve oxygen placed on the carbon by the HNO, treatment.ACFs with su rface areas and oxygen contents ranging from 700 to 2800 mZ/gan d 0 to 5 a%,espectively,were prepared commercially by reacting phenol forma ldehyd e fiber precurs ors (Kynol) in a s teadcarbondioxide mixture at temperatures between 700 and 900C. ACFs with acidic and basic surface cheniistlywere prepared u sing methods described elsewhere [8,9].Sample CharacterizationTh e SO, adsorption capacitiesof samples were determined by the rmog ravim etric analysis (Cahn TG -I 3 I).In a typical run, a 30-50 mg sam ple was placed in a platinum pan and heated at ZOWmin in flowing N,to 360C t o remov e moisture and impurities. The sample was cooled to 120C. Once the temperaturestabi lized, the N, was replaced by a mixtureofgas= containing 5% 0,, 7% H,O and the bala nce N,. Oncethere was no further weight gain due to adsorption of 0, and H,O, SO, was added in concentrationsrepresentativeofa typical flue gas from combustion of high su lh coal (2500 ppmv SO,). Th e weight gainwas recorded versus time by a computerized d ata acquisition system.Temperature programmed desorption (TPD) exp eriments were done in a flow-thru, 2.5 cm ID stainlesssteel fixed-bed reactor system. In a typical run, .5g of sample was heated in flowing nitrogen (0.5 Wmin)at 5"Umino a linal temperature of 1000C and held for 1 h to achieve nearly com plete desorption of C Oand CO, from the carbon surface. Non -dispersiv e infrared analyzers (Rosemo unt M odel 880) were usedto m onitor the concentrations of CO and C O , in the efiluent gas continuously.N, BET surface areas ofchars were de termine d using a sing le point BET adsorption equation with N, 77K) adsorption data obtained at a relative pressure (P/P.,) of 0.30 with a Monosorb flow apparatus(Qua ntach rom e Corporation). Surface areas were determined for ACFs with a Micromeritics analyzerusing nitrogen adsorption at 77K. Average pore sizes of ACFs were calculated using the Dubinin-Radushkevich equation and the nitrogen adsorption isotherm measured using a Coulter Om isorb .RESULTS AND DlSCUSSION

.

ih2lrAWThe results presented in Figure I suggest a lack of correlation between SO , adsorption and N, BET surfaceare a Tw o steam activated carbons prepared from IBC -IO2 with interm ediate surface areas (200 and 360

339


2/5

m'/g) and a commercial activated carbon, C algon F400 (1000 m2/g), adsorbed similar amou nts of SO2(between 15 and 20 w% S O 3 after about 4 hours. The Centaur carbon, which had the highest SO 2adsorption capacity, had an N, BET surface area of only 3 60 m2/g. T his carbon catalyst was developedby Calgon for both liquid and vapo r phase applicati ons, includ ing removal of SO , and NO, from flue gS .Figure 1 also shows that an activated carbon prepared from IBC -I02 coal by steam activation,mortream ent and subsequent heat treatment to 925C exhibitedSO, adsorption behavior similar to the Centaurcarbo n. One study [IO] showed that there was no correlation between the SO, adsorption capacity andsurface of a coal-based carbon, while anoth er [ I I] has maintained that "surface area is the mo st importantparameter in order to predict the behavior of a char in the abatem ent of SO , from exhaust gases."A better understanding of SO, adsorption behavior may require mo re detailed informationabout thecarbon-oxygen (C-0) complexes formed during char preparation and SO, adsorption. The authors showedpreviously a poor correlation between SO, adsorption capacity and the total amo unt o f chem isorbed oxygenfor cha rs prepared under a wide range of conditi ons [12-1 4]. Davini [I51 also observed no correlationbetween SO, adsorption and chemisorbed oxygen. He found a better correlation between SO, capac ity andthe basic (or high temperature) C - 0 func tional groups. In the current study, the steam activated char wasfound to have a surface populated predom inantly by high temperature C - 0 complexes: To increase thenumber of high temperature C - 0 complexes, and presumablySO, apacity, the H,O-activated char wastreated withHNO, nd thermally desorbed at temperatures ranging from 200 to 1075C. Figure 2 presentsthe TPD profiles of the HN0,-treated char and those thermally desorbed at 525 ,72 5 and 925C. The COand CO, evolutio n'profk of th e original HNO,-treated char showed only slight overlap . Con ceivably , thischar could be heated in inert gas to a certain temperature, e.g., 5 W C , to remove only the C0,-formingfunctional group s and retain the CO -formin g ones. Figure 3 shows that the SO, adsorbed increased withincreasing thermal desorption treatment. A three-fold increase in SO2 adsorption is observed with therelatively small increase in thermal desorption temperature from 525 to 725OC, and the char with thesmallest amount of C - 0 complex (925C) adsorbed the largest amount of SO,. This sug gests that sites thatrorm a stable C -0 complex during char preparation are made available by lh e therm al deso rption treatmen t,and that adsorption of SO, may preferentially occur at these free ad sorptio n sites. Using TPD , the au thorspreviously m easured the num ber of free adsorption sites for several carbons; a direct relationship wasfound between SO, capacity and free sites [12,14,16]. Based on these results, a reaction scheme wasrecently proposed IO explain SO, remov al by carb on whereby the free sites control the rate of adsorptionof SO, and conversion to H2SOI [ 7 ,l E].AmFigure 4 s how s that the SO, adsorbed varies inversely with surface area for the AC Fs studied. Th e porevolumes of these ACFs are known to increase from about 0.3 (ACF-IO) to 0.8 cm3/g (ACF-25) [19].Others have also observed a decrease in adsorption of SO, with increasing surface area and microporevolume of ACF (polyacrylonitrile fiber s) [4,6]. The oxygen and nitroge n(< 0.05%) content, although notmeasured directly for this series of ACF, should not have varied much amon g these four sampl es; althoug h,the oxygen content ofAC F-IO and 15 may have been slightly greater than t hat of A CF-2 0 and 25 [19,201.Exclu ding the idea of free sites for now, the results presented in Figure4 suggest that pore size was themost important factor for determining the SOzadsorption properties of these ACFs. These ACFs areknown to hav e relatively narrow pore size dis tributio ns [20,21]. In this stud y, we calculated the averagepore sizes ofA CF s 10, 15 , 20 and 25 to be 9.4& 11.7A, 13A and 17.5A. respectively. Foster et al. 1191studied a series of ACF similar to those used in this study and found that ror ppm levels of n-butane,benzene or aceto ne, low surfa ce area AC F adsorb ed mo re than high surface area ACF . It is well knownthat smaller pores are preferentially filled at low relative pressures (concentrations) of adsorbate due to theoverlap of attractive forces of opp osite pore walls. If a similar adsorption mechanism occurs in the SO,-ACF system as n the n-butane-ACF system, this would explain the behavior observ ed in Figure 4, and helpexplainthe ack of correlation betweenN,BET surface area and SO, capacity observed for the coal chars.Ullinialely, the num ber of rree sitesas well as pore size distribution and p ore volum e o r an acliva ted carbonshould d efine its SO, adsorption behavior. The relative importance of each remains to be determined.Figure 5 show s that the ACF-IO t reated with b oth sulfuric and nitric acids (H,SO&NO,) and thermallydesorbed to 40 0 and 700C exhibitsSO, adsorption behavior similar to that of the HN0,-treated, thermallydesorbed coal chars (Figure 3). 7h e SO2adsorption capacity of the H$OJHNO , treated carbon thermallydesorbed at 700C is much greater than thatofthe originalACF-10. The presence of adsorbed 0, on AC Fmay serve only to block access of SO2 to free active sites, similar to the behavior w e have ob served forIllinois coal char. Heating the HSO ,MN O,-treated ACF to 700C would gasify pnrt of the AC F, evolvi ngoxygenas CO or CO, thus increasing the average pore size. If the average pore size of AC F incre ases withoxidatiodthermal desorption, the results shown in Figure 4 w ould p redicta decrease in the SOzcapacityof H$OJHNO, treated ACF thermally desor bed at 700C. if pore si ze was the only cont roll ing fact or, Thefact that there was an increase in SO2 adsorption despite an increase in pore size suaests the stronginfluenceof Free sites in controlling SOzadsorption behavior. It remains to be determined whether the freesite concen tration of these ACFs can be mea sured. II is interesling thal h e hermal desorption ireatmentapplied to both the H2SO JHN 03-treat ed ACF and the HN 0,-treated coal char serves to increase not onlythe SO, adsorbed, but also the initial rate of SO, adsorption (Figures 3 and 5 ) ; this behavior has beenattributed to an increase in the n umber of active sites [22].

9

'

34 0


3/5

\

Modilying the surface chemistry of ACFs by incorporating different functional gro ups has b een shown toimprove adsorption of various adsorbates [8,9,19,20,23,24]. ACFs with both acidic and basic surfaces,and an average pore size smaller than the original ACF-IO, were prepared using methods described inprevious work [8,9]. Since SO, is an acid gas, a carbon with basic surface characteristics would beexpected to exhib it enhanced SO2adsorption. Indeed, Figure 6 shows that a basic AC F adsorbs more SOaIhan an acidic ACF. Others have also suggested that the presence of sm all am oun ts o f nitrogen i n ACF canmarkedly improve its SO, removal capabilities [4,6,25,26]. Work is in progress to learn how theinterrelationships amon g num ber o f free sites, pore size distribution and types o f functional groups maycontrol a carbons SO, dsorption behavior [27].CONCLUSIONSCoal-based activated carbons and ACFs, prepa red by a novel oxidatiodthermal desorption treatment, hadSOl adsorption capacities approaching that of a comm ercial carbon catalyst. Th e current price of ACFs,about $1 OAb., may limit their use in SO2 emoval applications, however, some recent work suggests thatACFs can be prepared for $ I - W b . [28]. Further work is needed to characterize these less expensive AC Fsystems for SOaremoval. Th e results obtained in this study suggest that pore size as well as the numberof free active sites play im portan t roles in determining a carbons SO, adsorption capacity. Th e completepore size distributions o rt h e treated coal-based carbons an d AC Fs remain to b e determined; these shouldprovide additional insight into the relative importance of pore size and free sites in controlling SOaadsorption behavior. Th e use of functional groups, e.g., nitrogen-containing ones, to mo dify a carbonssurface to impro ve SO, adsorption appears promising, and is also an are a for further investigation.ACKNOWLEDGMENTSThis work was sponsored by the Illinois Department of Energy and Natural Resonrces through th e Illinois CoalDeYeropment-Board nd Illinois Clean Coal Institute, and the United S tates Departm ent of Energy. Th e technicalassistance of Ms. Gwen Donnals is gratefully acknowledged. Discussions with Dr. M ark Rood, Dr. Carl Kruseand Dr. John Lytle are appreciated.REFERENCES1.2.3 .

4.5.67.8.9.IO.I I .

12.

13.14.15.16.17.18 .19.20.21 .22 .

Richter, E., Knoblauch, K., Jungten,H., 1987,GasSep. Pur$ I , p. 35.Tsuji, K., Shunishi, I. , 1991, Mitsui-BF Dry D esulfurizationand De nitrification Process using A ctivatedCoke, Proceedings of Ihc EPRl SO, Control Symposium, Washington, D .C., p. 307.Brueggendick, H., Pobl, F.G.,993, Operating Experience with STEAG s Activated Carbon Proce sses -dd tm - In European Waste Incinerator Plants, ProceedingsofTenth Annual International Pittsburgh CoalConference, p. 787.Mochida, I., Hirayama, T., Kisamori, S., Kawano, S., Fujitsu, H., 1992 , Lungmuir 8, p. 2290.Kisamori, S., Mochida, I., Fujitsu, H., 1994,Langmuir 10, p. 1241.Kim, J.Y., Hong, I ,Lee, J.G., 995, Influencesof Preparation Conditions of ACF on Catalytic Activityfor SO2Removal, Proceedings of 22nd Biennial Conference on Carbon, San Diego, CA , p. 534.Harvey,RD ., Kruse, C.W., 1988,J.Coal Qual.7,p. 109.Economy, J., FosIer, K, Andreopoulos, A. and Jung, H., 1992, Tailoring Carbon Fibers for AdsorbingVolatiles, Chem Tech, p . 597.Em ~ ~m y ,., Daley, M.A. and Mangun, C.L., 1996, Activated Carbon Fibers - Past, Present and Future,ACS Preprints, Fuel Chem. D iv., New Orleans, LADavini, P., 1989,Fuel68, p. 145.Rubio, E., Izquierdo, M.T., Mastral, A.M. and Mayoral, C., 1994, Proceedings of InternationalConference on Carbon, Use of Low Rank Coal Chars in the Abatement of SOa from Flue Gases,Granada, Spain , p. 402.Lizzio,A.A., DeBarr, J.A., Donnals,G.L., e , C.W., Rood, M.J., angwal, S.K., 1995,Productionand Use of ActivatedChar for Combined SOJNO, Removal, Final Technical Report, Illinois Clean CoalInstitute, Carterville, IL.DeBarr, I.A. and Lizzio, A.A., Production of Activated Char for Combined SOf lO , Removal,Proceedings of International Conference on Carbon, Granada, Spain, 1994, . 268.Lizzio, A.A. and DeBam, J.A., EMect of Surface Area and Chemisorbed Oxygen on SO, AdsorptionCapacity of Activated Char, Fuel 1996 (in press).Davini, P., 1990, Carbon 28, p. 565.DeBarr, J.A., 1995, The Role of Free Sites in the Removal of SO, from Simulated Flue Gases byActivated Char, M .S. Thesis, Universityof Illinois, Urbana, IL.DeBarr, J.A. and Lizzio, A.A., 1995, New Insights on the Mechanism of SO, Removal by Carbon,Proceedingsof 22nd Biennial Conference on C arbon, San Diego, CA , 1995, p. 562.Lizzio, A.A., DeBan, J.A., 1996, The Mechanism of SO, R emoval by Carbon, ACS P reprints, FuelCheni. Div., New Orlcans, LA.Foster, K.L., Fuerman, R.G., Economy, J., Larson, S.M., Rood, M.J.. 1992, Chem.Mater. 4, p. 1068.Cal. M.P., 1995, Characlerimlion or Gas Phase Adsorplion Capacily of Untreated and ChemicallyTreated Activated Carbon C loths, Ph.D. Thesis, University of Illinois, Urbana, 1L.Lin, R Y ., Economy,J., 1973, AppliedPolymer Symposium 21, p. 143.Jungten, H. and Kuhl, H., 1989,Chem. Phys. Carbon 22, p. 145.

341


4/5

23 .

24 .2s .26 .21 .28 .

.-i- 250020003 15001000500 -0

Cal, M.P., Dimotakis, E.D., Economy, J., Larson, S.M.nd Rood,M.J., 1996, The Effect ofChem cialModification of ActivatedCarbonClothon he Adsorption Capacity ofOrganics and Water Vapor, ACSPreprints, Fuel Chem. Div.. New Orleans, LA.Daley, M.A., Mangun, C.L., Economy, J., 1996, Predicting Adsorption Properties for ACFs, ACSPreprints, Fuel Chem. Div.,New Orleans, LA.Nishijima, A., Kurita, M., K i y o d , Y., Kobayashi,R, Hagiwara, H., Ueno, A,, Sato, T., Todo, N., 1980,The Chemical Socie v ofJ ap m 53 , p. 3356.Fei, Y., Sun, Y.N., Givens, E., Derbyshire,F., 1995, Continuous Removal of Sulfur Oxides at AmbientTemperature, sing Activated Carbon Fib ers and Particulates, ACS Prep rints, Fuel Chem. Div.,40 (4),p. 1051.DeBarr, J.A., 1996, Ph.D. Thesis, University of Illinois, Urbana, IL.Daley, M., 1996 , Ph.D. Thesis, Universit yof Illinois,Urbana, IL.

HNO3-725C HNO3-925C--

co~ /

350

FigureTim e (h)

1. Effect of N 2 BET urface areea on SO 2 adsorptionIBC-102 chars and commercial carbons. capacity of selected

0

342


5/5

3501ACF-IO, 600m'/g

3~~ ACF-IS, 1400m'/gACF-20,1600 m'/gI

0 2 4 6 8 1 0 0 2 4 6 8 1 0Time (h) Time (h)

Figure 3. SO2 adsorption for HN 03-trea tedchars desorbed to different temperatures.Figure 4 . SO2 adsorption for A CFs

0 2 4 6 8 1 0 0 2 4 6 8 1 0Time (h) Time (h)

Figure 5 . SO2 adsorption for oxidized A CFsdesorbed to different temperatures. Fi gu re6 SO2 adsorption for acidic and basic ACE

343

Documents

Adsorption of So2 on Bituminous Coal Char