Document wr8rvLRJqLKMv8K5kYLyLOkeo

Internal Correspondence CU*.r:--* v------- 7 To R. M. Pieper From j. a. Kvikstad Subject Trichloroethylene Date January 7, 1986 I have attached a copy of an article which appeared in a recent issue of American Industrial Hygiene Association Journal, discussing the effects of relative humidity on adsorption of trichloroethylene onto activated carbon. Since we have seen increased mention of trichloroethylene in the literature, I would like to suggest that we examine the capacity of the #3500 OVM for this compound, and how that capacity is effected by humidity. JAKismt Attachment c: A. G. Holcomb R. E. King J. B. Palazzotto A* f ypc - - 3H 006547 All, ll.U H\^ I 41,i 1i r lw> The Effects of Relative Humidity on the Vapor Phase Adsorption v .Trichloroethylene by Activated Carbon MAK I IN U ttlKMK l nivvrsiiy el 1 t:\iis, Ncheel ol I'uhlii Hc.ilih The effect of relative humidity on activated carbon adsorption of trichloroethylene (T('t) Mas determined. Five levels of relative hun.i, were tested for each of four TC'E influent concentrations. The effect of humidity was greatest at the lowest TC'K influent concentration mg/m ) and highest relative humidity involved (859c ) l nder those conditions the amount of TC t adsorbed was only 99c of the amm. adsorbed at the same challenge concentrations and lowest relative humidity (59c). The effect of humidity was negligible at the highest I i concentrations (1000 and 1300 mg/m) and moderate level of relative humidity (259c)- l nder all other conditions tested the effect intermediate to these extremes. Differences in TCE breakthrough curves caused by the presence of water vapor are presented and discussi Data sets (grouped by relative humidity) fit the Dubinin - Polanyi equation equally well. Introduction Activated carbon adsorption has numerous applications which involve the removal of organic compounds from air. including. (1) occupational respirators.'1 '' (2) air pollution control devices."'r (3) air sampling dev ices.111 and (4i vapor recovery operations(IJ The use ol activated carbon is. in general, relatively expensive and can involve cases lor which process failure can lead to direct health risks (eg.. occupational respirators). Therefore, reliable criteria for process design and operation are highly desirable. Although ns impact is poorly understood, relative hum . ity is an important lactor w hich may adversely allect ad',: tion processes.'' u>Since various humidity leveis e.i during each of the applications of activated carbon tv,, tioned. relative humidity should be considered during dt.'.r and application of many types ol activated carbon system' Therefore, the purposes ol this paper are: (I) to prese; results which illustrate the degree to which several leveis o relative humidity reduced the adsorption capacity of acti 1. DRY, PURE AIR GENERATOR 8. HUMIDITY GENERATOR 2. PRESSURE VALVE S. EQUALIZATION VESSEL 3. AIR FLOW CONTROLLER 10. TEMP PROBE 4. IMPINGER TUBE FOR TCE 11. DEWPOINT SENSOR GENERATION 12. HYGROCOMPUTER 5. CONSTANT TEMP BATH 13. PRESSURE GAGE 6. ROTOMETER 14. ACTIVATED CARBON COLUMN 7. FLOW VALVE 15. GAS CHROMATOGRAPH Figure 1--Diagram of experimental system Am Inti Hyg Assoc J (461 October 198S 3M 006548 5*b mental tiul' tiquueJ. luMi. oii.-ti.. '<. u...- ; ,, plctc. depending mi the spciiiic ionditu>ii' of 1C 1 e mw fil tration and humidity. Determination of adsorption capacity wete based on time to 50f,( TCE breakthrough aqd the assumption ol a sy mmet ric breakthrough curve.'' ' ns Results and Discussion Ellacts of Humidify on Adsorption Capacity The quantities of TCE adsorbed per gram ol actuated car bon are presented m Table I lor the individual trials. Eor similar inlluent TCL concentrations, the amount of TCE adsorbed decreased with increasing relative humidity. Although other investigators have noted the adverse effect of humiditv on gaseous phase adsorption.'1IJ u data were collected during this research in a systematic manner specifi cally designed to permit a detailed analysis ol the impact of relative humidity, The ellect of increasing relative humidity on the amount ol TCE adsorbed tor each inlluent TCE concentration is illustrated in Figure 2. Additionally, an analysis of similar results from other studies is summarized in Table 11. The ellect ol humidity was more pronounced at the lower TCE influent concentrations tested than at higher concentrations. The magnitude ol the humidity ellect in the present study is comparable to results reported by two investigators.u but much greater than results reported by another (Table 11 At least lour factors influence the impact of humidity on gaseous phase adsorption, however, and should therefore be considered when comparing studies. First, the manner in which carbon is preconditioned often differs between stu dies. For example, one investigatorIM1 preconditioned the carbon in an atmosphere of 509r relative humidity; in another study"1 preconditioning procedures varied between trials; the third investigator114'apparently performed no pre conditioning; and this author preconditioned the carbon by oven drying for two days. Since the manner in which the RELATIVE HUMIDITY (*) Figure 2--Percent of TCE adsorbed at each level of hum c ; relative to that adsorbed at the lowest humid i> carbon is preconditioned has significant effects on thi adsorption efficiency of carbon,u;i' a complete quantitative comparison between the studies (as represented in Table II is tenuous. The second factor which may influence the impact o humidity is adsorbate concentration. Results of this stud; indicate that low concentrations of adsorbate are affected b; a given level of humidity to a greater extent than highe concentrations. Thus, direct quantitative comparison TABLE II Results from Selected Studies on the Effect of Humidity on Carbon Adsorption Under Various Experimental Conditions Adsorbate influent Adsorbate Reference Teat tk Adsorption Cone, (ppm) Humidity % Humidity Capacity Retained Reference Benzene 1-Chlorobutane Carbon tetra- chloride Ethyl acetate Acetone Hexane Benzene Chloroform Acrylonitrile Chloromethyl methyl ether Acrylonitrile T richloroethylene Trichloroethylene 1000 1000 1000 1000 1000 1000 500 500 100 1 10 222 50 20 20 20 20 20 * 0 15 15 15 15 7.5 5 5 80 80 80 90 90 90 80 80 80 80 90 85 85 86 13 79 13 78 13 77 13 87 13 83 13 73 4 67 4 45 4 30 4 17 14 28 Present study 9 Present study Am ind Aso: J (46; October mb 3M 006549 SB TABLE I TCE Adsorbed (or Individual Trials Relative Humidity TCE Influent Concentration TCE Adsorbed <*> mg/m m ppm (g TCE/g carbon) 5 303 52 0 286 25 295 50 0 257 SO 293 50 0 1B0 65 295 50 0 114 85 293 50 0.027 5 602 103 0 334 25 605 103 0.320 50 597 102 0 284 65 599 102 0 160 85 593 101 0 054 5 987 168 0 399 25 995 170 0 403 50 978 167 0.342 65 996 170 0218 85 986 168 0 098 5 1331 227 0.434 25 1306 223 0.431 50 1356 231 0 370 65 1322 226 0 262 85 1304 222 0.121 'Calculations were based on time to SO'to TCE breakthrough jted carbon for a single volatile organic compound (VOC) . iovv influent concentrations. (2) to examine the interaction /tween VOC concentration and humidity during carbon dsorption and (3) to suggest a potential approach that at be taken for predicting the degree of impact of relative jrmditv on the adsorption process. xperimental Methods and Design he experimental system employed is illustrated in Figure I. >rv t5'r or less relative humidity) and purified air entered :e system through a pressure valve (2) (numbers in parentcsex refer to system components in Figure I). The air was .en immediately split into two streams. One stream (0.2 .5 ( of the total flow) was directed through a mass flow unsducer (3) to an organic vapor generating apparatus, or these experiments, the vapor generating apparatus con-;ed of a 4.5 cm impinger tube (4) partially filled with ichloroethvlene (TCE) in a temperature controlled water -i t h (51. 1 he concentration of TCE in the air was controlled ail ot the following mechanism: regulating air flow to the "crating apparatus, regulating the temperature in the -iter bath and setting the distance between the air entry nt in the impinger tube and the surface level of TCE. ! he second air stream (10 L; min) flowed through a ometcr (6) and into the main portion of the system. This : stream also could be split to keep any portion of the total 'w dry. while the remainder could be humidified by expo* . water. The air humidifying system consisted of two - :hes connected in series (S). Each tube was 5 cm in :er and was partially tilled with water. I he tubes con- .ola'nc pall rings which served to disperse the air and ..:*c ai,-water contact. The three portions of air (TCE laden, water vapor laden and pure dry air) were then united and directed through a 20-liter equalization vessel (9) to dampen short-term varia tions in TCE concentration and air humidity. The total volume of air then passed over a temperature probe (10). A low air How (25 ml min) entered a dew point sensor (11). Signals from the temperature probe and dew point sensor were continuously relayed to a hygrocomputer (12) to per mit monitoring of dew point and relative humidity. The air stream was again divided: 2.3 L. min was diverted to a gas chromatograph for influent TCE concentration determina tion. while the re-maining 7.7 L min was directed through the activated carbon column (14). The effluent from the column was also directed to the gas chromatograph for analysis. A pressure driven solenoid valve mounted on the gas chromatograph permitted alternating samples of the influent and effluent air to be taken on a programmed time schedule. The time interval for a complete sampling cycle during the experiments was 20 minutes. The flow (7.7 L, min of test air) through a 2.54 cm diame ter activated carbon column resulted in a linear velocity of 25 cm, sec and an air retention time of 0.5 sec within the carbon. For all experimental trials the mass of activated carbon used (37.5 g of C-ECA G AC48C Carbonundum) was constant and resulted in a carbon bed depth of approximately 13.5 cm. The carbon was washed with distilled water and oven dried for a minimum of two days. Experimental trials were performed at four TCE concen trations (approximately 300. 60(f, 1000 and 1300 mg/m\ each at five levels of relative humidity (5.25.50.65 and 85c*). Temperature levels were maintained at 232C. Single trials were conducted for each experimental condition. Experi- 3H 006550 J i 'Stef TABLE III Time to Broakthr ugh of 50% of th Influ ent TCE C ncentration Standardized to that Valua at Low Ralativa Humidity (S^J. Ralativa Humidity TCE Influent 5 25 50 65 85 Cone, (mg/m') Ralativa Amount Adsorbed 300 600 1000 1300 100 90 63 40 9 100 96 85 48 16 100 101 87 55 25 100 99 85 61 28 etween adsorbate concentrations are invalid. Related to his point, the "multiplier factor" suggested as a tool to ccount for the effect of humidity on activated carbon respiators'11' is probably only reliable within a range of adsor bate concentrations near the concentration tested. Specifi.ally, a "multiplier lactor" should not be used directly for -hallenging adsorbate concentrations that are much lower than concentrations for which the factor was derived. The third factor which influences the impact of humidity is the type of adsorbate compound. The degree to which humidity affects adsorption is a function of the adsorbate.14' Specifically, hydrophobic adsorbates are expected to be more adversely affected by humidity than hydrophilic compounds.'11 The fourth influential factor is the type of actuated car bon. The carbon structure can influence the amount of water sapor adsorbed as well as the competitive interaction between water molecules and organic adsorbate molecules lor available sites on the carbon. For example, one study demonstrated that for several types of activated carbon the o \y gen content of the carbon w as positiv ely correlated to the carbon's affinity for water molecules."" Since humidity can have substantially adverse etfects on the gaseous phase carbon adsorption process, humidity should be considered during the establishment of design and opera tional criteria. But. us argued, quantitative statements based -vilely on laboratory results which investigate the impact of numidity ot specific compounds under a given set of condi tion- can not always be generalized to other situations. The extent to which results of research, such as that reported here, are general for other adsorbates, adsorbate concentra tions. and activated carbon types needs to be determined by additional experimentation under similar conditions. Ulti mately. research is needed to assess mechanisms by which humidity interferes with adsorption lor various classes of 'i game compounds and adsorbate concentrations. Then pieJictive equations could be developed to extrapolate experimental results to a large number of combinations of conditions which exist under actual operating conditions. Shape of the Breakthrough Curve \ ic-cmh introduced parameter.'- fl termed the relative tund.iid dev union (or), desci ihes the shape ot an adsorption ''ic.iK'.hiough curve. i. = .1 t TABLE IV Time to Breakthrough of 10% t the Influ ent TCE Concentration Standardized to that Value at Low Relative Humidity (SS). . Relative Humidity TCE Influent 5 25 SO 65 65 Cone, (mg/m ) Relative Amount Adeorbed 300 600 1000 1300 100 67 58 32 9 100 62 77 38 15 100 93 80 48 24 100 87 83 51 27 Where: o= Standard deviation of the adsorbate breakthrough curve. (The calculation of a is based on tabulations of the normal probability distribution and the assump tion of a symmetrical adsorbate breakthrough curve about the 50Cr breakthrough point).,J1 Tso= Time to 50f r adsorbate breakthrough. This parameter should be independent of adsorbate chal lenge concentration. Broader contaminant breakthrough curves have greater values of or than do narrower curves. The mean value of or for the low humidity (5(<) trials of this research was 0.126 (range 0.100 to 0.167): for all higher humidity trials or was0.210(range0.156 to 0.298). Thus, the presence of water vapor caused the shape ol the TCE break through curve to be broader than would have been predicted by results at the lowest relative humidity tested (5C<). Values in Tables ill and IV are length of the time for all experimental trials to 50('i and I0r( TCE breakthrough, respectively. Each value is standardized to the length ot time to breakthrough for the corresponding low humidity trial. (Note, the standardized times at 5(K< breakthrough are comparable to the standardized quantities of TCE adsorbed, since the midpoint of the breakthrough curve was used to calculate adsorptive capacity).'-' If relative humidity had no ellect on the shape of the breakthrough curve, values in Tables III and IV would be identical. Dilferences in the values reflect the effect ol water vapor(i.e.. boadeningof the TCE breakthrough curve). Practical implications of the effect of relative humidity on the shape of a breakthrough curve depend upon the applica tion of the adsorption process. Broadening of the curve has little importance for air pollution control processes in which carbon columns are used in series. For those processes the upstream column can remain in service until its carbon is completely exhausted. However, a broadening of the curve would be significant for operations in which a single carbon bed is-involved occupational respirators). Under con ditions of high humidity and w hen the threshold limit of the contaminant is below 50(r ol the ambient concentration, the unit may have to be removed from service sooner than would be predicted by the carbon's adsorption capacity and the shape of the breakthrough curve obtained at a lower relative humidity. Research on other compounds at various relative humidities is required to determine it the ellect ol water vapor on the shape ol the ICE breakthrough curve is a common phenomenon. 3M 006551 A Potential Predictive Approach 'I he I)uhimn-IVIunv i isniheim cquutuui in it' linear i/ni form follow s' In W = In Wo - U>. /3') (R I In (I'o l'>)` Where: W = volume ol adsorbate which i> adsorbed imL adsot- bate, g adsorbent). Wo = limiting adsorption volume (ml g adsorbent). * - parameter for each adsorbent-adsorbate svstem. /? = aflimtv coefliciem ol adsorbate, R = gas constant 00143 J K mole). T = temperature (k). Po = saturated vapor pressure ol adsorbate at a given temperature. P = actual vapor pressure of adsorbate under the specific operation conditions. The equation is based on rational chemical and physical principles and was specifically designed to describe adsorp tion ol gaseous compounds on porous surfaces.'16 ' The equation has been show n to accurate!) fit experimental data and has proven to be a uselul predictive equation for adsorp tion in single adsorbate systems.'1" Recent advances have been made in predicting parametei values of the equation DUBININ-POLANYI ISOTHERM Figure 3--Experimental results fit to the Dubinin-Polanyi iso therm equation 4m Ina Hyg Aaoc J (46j October 1985 t ion to: aCi U.O appiita! ion- Ii.: m. '.: pteo ic'.'O' - . on phv sicochen.it.il pniptiins .u.c etnpiris.. re'.. - I o Ini iIn i an.ilv /e the effect ol water vapor on :n. ..c - lion process, each set ol data corresponding to a sp.i level ol relative humiditv was In to the Dubimn-Polat equation Result' ate shown in bigure 3 Based on ti results, tw o muioi poinis can be m.uk hirst, each >e, , . tits the equation equally well indicaline ;:,. cue h humid it v lev el the el led ol vv u let v a no: wa - ci'i> s w ith the theoteiical basis ol the Dubmin-Polanv i cun. ' Thus, to predict the died ol relative humiduv on earn, adsorption, one only needs to determine how the parurnd. values. (3 and s>. aie altered by the presence ol watci vjp, lor a given adsorhate-adsorhent svstem 1 he solubilitv o! ; i. adsorbate in vvutet. along with other physical ehen,:... properties ol'tlle adsorbate and adsorbent, mlluciie.. died ol humiduv on the adsorption pi oeess; these pi ope11 . - may result in predictable changes m the values ol p oi s.. c both. Notice, the lines in Figure 3 converge as values alou_ the abscissa decrease. 7 hus. the value ol \\ o (the 3 intcreep' in Figure 3) wus not appreciably changed at dilterent relatu . huniidit> level' Apparently, the concentration ol wau vapor had little ellect on the limiting adsorption v olurn. o the carbon lor 1 CL. Second, the trend seen in Figure 3 clearly illustrates an important interaction between the level ol humiduv and adsorbate (TCE) concent rat ion. (I nlluent adsorbate concen trations increase as values along the abscissa decrease.) The deleterious impact of humidity is less at high TCE concen trations than at lower concentrations. In fact, if the adsor bate concentration were extended beyond the experimental range, the lines in Figure 3 would meet. Thus, based on an extrapolation of these data, one could predict that at some high adsorbate concentration the impact of humidity would be negligible. Experimental results substantiate this hypothesis; at the two highest TCE concentrations (1000 and 1300 mg m3). 259r relative humidity did not reduce the amount of TCE adsorbed (Table 1). Based on these results, the relatively small impact of humidity reported in one study113' (Table II) can largely be explained by the high influent adsorbate concentration used (i.e., four times the highest adsorbate concentration used in this research). If results were extrapolated to very low organic adsorbate concentrations and high relative humidity, carbon adsorp tion could become impractical, based on the trends demon strated in Figure 3. Conclusion Relative humidity at five different levels affected the adsorp tion capacity of activated carbon. In general, each increase of humiditv level further decreased the carbon's adsorptive capacity at all four TCE concentrations tested. Further more. at low TCE concentrations the effect of humidity was greater than at high TCE concentrations. The results indicate that two assumptions sometimes made concerning the effect of humidity on vapor phase 3M 006552 actuated carbon adsorption should be applied only with caution. First, making an assumption that relative humidi ties below 5(Kf have negligible impact on adsorption may be invalid. All levels of relative humidity had an adverse effect at the low TCE concentrations tested during this experi ment. Second, applying a "multiplier factor" determined at one adsorbate concentration to account fora given humidity at other adsorbate concentrations is tenuous. As indicated previously, the effect of humidity on the carbon's adsorption capacity was strongly influenced by the influent TCE con centration. Additional research is required for other adsor bates under similar experimental design to determine how general these results are. The Dubinin-Polanyi isotherm equation fits data at each humidity level equally well. The basis of the equation may be useful for predicting the effect of humidity on gaseous phase carbon adsorption. Determining the mechanism by which the values of the equation parameters (specifically, /J and *s) are influenced by humidity will be critical in applying the predictive power of the equation to the impact caused by humidity. References 1. Nelson, G.O, and A.N. Correia: Respirator Cartridge Effi ciency Studies. VIII Summary and Conclusions. Am. Ind. Hyg. Assoc. J. 37:514-525 (1976). 2. Grubner, O. and W. A. Burgess: Simplified Description of Adsorption Breakthrough Curves in Air Cleaning and Sam pling Devices. Am. Ind. Hyg. Assoc. J. 40 169-179 (1979). 3. Grubner, O. and W.A. Burgess: Calculation of Adsorption Breakthrough Curves in Air Cleaning and Sampling Devi ces. Environ. Sci. Techno/. J5:1346-1351 (1981). 4. Stampfer, J.F.: Respirator Canister Evaluation for Nine Selected Organic Vapors, Am. Ind. Hyg. Assoc. J. 43.319328 (1982) 5. Moyer, E.S.: Review of Influential Factors Affecting the Per formance of Organic Vapor Air-Purifying Respirator Car tridges. Am. Ind. Hyg. Assoc. J. 44:46-51 (1983). 6. Turk, A.: Source Control by Gas-Solid Adsorption and Related Processes. In Air Pollution, A.C. Stern, ed pp. 479519, Academic Press, New York (1968). Schabe, P.: Gas anc vapc, Fi'trat'C'1 w.:n Speoa' Re'e-' ence to Activated CarDon Fiiit Sec 8 153-157 p9'i > 8 Ooig. I.: Active Carbon m Air Pollution ana Odour Contro Clean Air (Brighton, Engl.) 14.55-62 (t980) 9 Howard, A.G. and R. Pizzie: Sampling of Chlorinated anc Brommateo Haliforms from Humid Atmospheres Ann Occup Hyg 24 167-173 (1981) 10 Gregory, E.D. and V.J. Elia: Sample Retentivity Properties of Passive Organic Vapor Samplers ano Charcoal Tubes Under Various Conditions of Sample Loading, Relative Humidity. Zero Exposure Level PerioOs and a Competitive Solvent Am. Ind. Hyg. Assoc J 44 6B-96 (1983) 11. Underhill. D.W.: Efficiency of Passive Sampling by Adsor bents. Am. Ind Hyg. Assoc J. 45 306-310 (1984). 12. Pamele, C., W. O Connel, and H. Basdekit: Vapor Phase Adsorption Cuts Pollution, Recovers Solvents Chemical Engineering 1979: 58-70 (1979). 13. Nelson, G.O., A.N. Correia and C.A. Harder: Respirator Cartridge Efficiency Studies VII Effect ot Relative Humid ity and Temperature Am. Ind Hyg, Assoc. J. 37.280-288 (1976) 14 Henry, N.W. and R.S. Wilhelme: An Evaluation of Respirator Canisters to Acrylonitrile Vapors Am. Ind. Hyg. Assoc. J. 40:1017-1022 (1979). 15. McDermot, H.L. and J.C. Arnell: Charcoal Sorption Studies II. The Sorption ot Water by Hydrogen-Treated Charcoals. J. Phys. Chem. 58 492-498 (1954) 16. Cerny, S,: Theory Adsorplion on Activate Carbon on Active Carbon. In Active Carbon- Manufacturer Properties, and Application. M Smisek and S. Cerny. eds pp. 71-161 Else vier Publishing Co., New York (1970) 17 Dubinin, M.M.: The Potential Theory of Adsorption of Gases and Vapors for Adsorbents with Energetically Nonuniform Surfaces Chem. Rev 60 235-241 (i960). 18. Urano, K., S. Omari and E. Yamamoto: Predictive Methods for Adsorption Capacities of Commercial Activated Car bons in Removal of Organic Vapors. Environ. So. Technoi. 76:10-14 (1982). 19. Sanone, E.B., Y.P. Tewarf and L.A. Jonas: Prediction of Removal of Vapors from Air by Adsorption on Activated Carbon, Environ. Sci. Technoi. 73.1511-1513 (1979). 20. Sansona, E.B. and L.A. Jonas: Prediction of Activated Car bon Performance for Carcinogenic Vapors. Am. Ind. Hyg. Assoc. J. 42:688-691 (1981). l-l Dewmber 1484: KtfMsvd 74 April 1485 3M 006553