7/27/2019 julito4

1/5

J. Chem. Eng. Date 1992, 37, 339-343 339Equilibrium Constants for Methyl tert-Butyl Ether Liquid-PhaseSynthesisJ& Fdlp. Izqulerd o, Fldel Cunlll, Merkxell Vlla, Javler TeJero,and Montw rrat Ibo rraChsmical EnginWng Depertment, Unhwslty of Barcelona. Martii FranqGs 7, 08028 Ba rc eha , Spain

Equlltbrlum c onstants for the Ilquld-phase syntherlr ofmethyl 1.rt-butyl ether (MTBE) were determinedexpwhenta l ly In the temperature r a g . 40-80 OC.EquIWbrlum was establI8hed In the methano l addltlon tohbutyhe or obtalnhg MTBE over the rulfonk acid rerlnLewatlt 8% 118 a8 he catalyrt In a batch reactormalntatned at 1.6 MPa. The experh nental equlllbrlumconstant k glven as a functlon of temperature, andenthalpy, free energy, and entropy changes have beendetermlned as well. At 298 K, A H o =-38.0 f 0.8 kJmol-', AGO =-13.6 f .8 kJ mol-', and A S o =-81.7 f0.5 J mol-' K-'. The UNIFAC ert lmates of actlvltycoWkhta have been wed to dercrlbe the Ilquid-pharemmkkal l ty . A comparkon of expwbnental,th.rmodvn amlc, and llterature data has beon carrled out.Introductlon

There is currently great interest in methyl tert-butyl ether(MTBE) because of its excellent antiknock and poisoningemission reduction properties. The MTBE production is ex-pected to be doubled in a few years.MTBE Is obtained by the addition reaction of methanol toisobutene. The reaction is reversible, moderately exothermic,and usually catalyzed bymacroporous sulfonic ion resins (mainlyAmberlyst 15 and Lewatit SPC 118). The selectivity is ex-tremely high, but some byproducts such as dimethyl ether anddilsobutylene can appear if the temperature is hlgh enough andthe molar methanoi/isobutyiene ratio is far from the stoichio-metric one ( 7 ) . The presence of tert-butyl alcohol is alsopossible if the reactor feed contains water.Despite the fact that more than 1000 papers about MTBEcan be found in the literature, only a very few of them bring outthe reaction thermodynamics ( 7 - 4 ) . Besides, as it will bepresented inthe Results and Discussion, the dispersion of datais significant in such a way that itwould be a big problem tochoose the better one.The aim of the present work is to determine experimentalvakreg of the equlibriumconstant bydirect measurement of themixture composition at equilibrium and to compare them withthosecalculated from thermodynamic data and those found inthe open literature. The temperature range covers 40-50 OC,for which we did not flnd experimental equilibrium constants.Experlmontal Sectlon

Matd8b. Methanol HPLC (ROMIL Chemicals), with a min-imum purity of 99.8% containing less than 0.2% water, andMTBE (MERCK, Schuchard) with a minimum purity of 99%containing less than0.1 wt % water were used. Isobutyleneof 99% purity was supplied by SEO, Barcelona, the main im-purltles being isobutane and linear butenes which do not reactunder our reaction condltlons, and used without further purifi-cation. Nitrogen supplied by SEO, Barcelona, with a minimumpurity of 99.998% was used to achieve the suitable pressureto maintainthe reacting mixture in the liquid phase.The ion-exchange resin LewatitSPC 118 BO (now K-2631)(Bayer), used as the catalyst, is a macroporous sulfonated c+polymer of styrene-dlvlnyibenzene (DVB) contalnlng a matrix

cross-linked wlth approximately 18 % DVB with a surface area(BET method) of 36 f 1 m2-g-' and an exchange capacitydetermined by titration of 4.83 mequlv of HS03*g-' of dry resin.The beadsize distribution ranges from 0.25 to 1.6 mm, and theeffective bead size is 0.63 mm(f0.05 mm).A#nw.tur. Theexperiments were performed in a stainlesssteel jacketed reactor (200 cm3) in a batchwlse operation.Figure1 shows the experimental setup. Thetemperature wasmeasured with a thermocouple submerged in the liquid phaseand was controlled within f0.2 OC by thermostated water(thermostatic bath, LAUDA s15/12) containing propylenegtycolthat flows through the jacket. The reactor was connecteddirectly to the liquid sampling valve (VALCO 4-CL4WE) whichinjects a small pressurized liquid volume (0.2 pL) to the gaschromatograph (HP 5890A). This technlque has been shownto be suitable to analyze liquids and liquifiedgases with anexcellent reproducibility (5).A n a m . Helium(SEO, Barcelona) with a minimum purityof 99.998% was usedto helpthe liquidsample tovaporize and,at the same tlme, to carry itto the conductivity sensor at aRowrate of 30 cm3min-'. A 3 mX 3.2 mm0.d. stainless steel U Ccolumnpackedwlth Chromosorb 101 (80/100 mesh) was usedto separate the mixture of MTBE, methanol, isobutylene, andnitrogen. The column was temperature programmed with a5-min initial hold at 110 OC followed by a 10 OC/mn ramp upto 220 OC, and held for 5 min.

proM

7/27/2019 julito4

2/5

340 Journal of C h o m h l and Engineering De&, Vol. 37,No.3, 1002

I I -T-Zi"K i l

Flgurr 1. Experlmental setup.temperature was lower. The number of samples was as smallas possible in order to reduce and avoid compound losses,particularly the sobutylene, everytime thatthe reactor pressurewas restored.(7) I f the amount of byproducts was negligible, the samecharge was usedto study theequ#brkwn atothertemperatures.On t he contrary, a fresh mixture was replaced.(8) At lower temperatures (40 and 50 'C ) a MTBE amountwas initially charged into the reactor to lower the waiting timeto reach t heequilibrium. Even under that strategy the waitingtime lasted sometimes 8 or more h.(9) Experimental equilibrium constantK values were com-puted from the composition by using the following expression:

I t mustbe recognized that only the reacting species have tobeconsidered for the K caiculatbn, but ail the species presentin solution, both reacting and inert, have to be taken into ac-count to calculate K,Computation of the EqulUbrlumConstant fromTharmochemlcal Datareaction of a nonideai system is given by

Kx =x M l B E / x d / 4 4 H 8 (1)

The thermodynamic equilibrium constant for a liquid-phase

(2)and can be related to the thermodynamic properties of thesystem through the equation

The standard free energy change for the iiqulbphase reactloncan be computed from the standard enthalpy and entropychanges:(4 )

The standard entropy change for the liquid-phase reactioncan be deduced from the vaporization heats through the ex-pression

K =exp( -AG' , ) /RT) (3)

AGO,,)=AH',,) - TAS', ,)SAS'( ,) (T)= - ~ / T ) C V , A H ' , , ( T ) (5)/=1

where each AH',, is obtained by theWatson relation (eq 6):AHOAT) =A Hov (T oX ( l- T/Tc) / ( l - To/T , ) )0 .38 (6)

Similarly, the reaction enthalpy change is given bySI 1AH',,, =AH',,) - ~ v , A H ' , , ( T ) (7 )

Table I. Thermochemical Data of Methanol, Isobutene, andMTBE (Standard State, Liquid at 1 atm and 298 K) ndConstants for the Equation CJ(J mol-' K-') =a + b ( T / K )+c(T/KV+d(T/KVmethanol isobutene MTBE

A H 0 , / ( J mol-I) 37940 20 660 30 3800 1 1391.6O 596.9 53.18bc, x 102 3.781' 1. 44 P -0).1533bTc/K 512.6 417.9 496.4bi -12.36' -4.638' 0 . n ~d , X lo3 -3. 719 -1.372' -0.202 4b

'Calculated from the Gallant fitting of a third-order equation(16). bEs tima ted by the R owlinaon equation (17).By introducing eqs 5 and 7 in eqs 3 and 4, we can finally getthe dependsnceof AGO,,,and K on temperature.The dependence of AH',,, on temperature can also becomputed by integrationof the Klrchoff equation (eq 7):

S/=1dAH'(,)/dT =CvF,( ,)r (8)

where CPmare t hemolar heat capacities in the iiquld phase ofthe species that take part In the reaction and are usually ex-pressed in the polynomial form(9)

By integrating eq 8 and taking into account eq 9, we obtainCP(,)/ a,+b,T + c,T2+d/T3

AH',,, =1 , +8T +( b / 2 ) T 2+(C/3)T3 +(d/4)T4 (10)where

s S Si=1 /=1 /=1a =C v , a, b =C V ,, c =C v , ,

P (1 1)d =c v , d,

/=1

The integrationof the van't Hoff equation (eq 7)d in K/dT =A H o / R T 2 (12)

considering eq 10 givesin K =ZH - I K / R T+( a / R ) n T +

( b / 2 ) R T + (c/6)RT2 +(d/12)RT3 (13)The constants I, and I, can be calculated from the temper-ature dependence relationship for the enthalpy change of re-action and for the equilibrium constant, respectively.The thermochemical data required for evaluation of equiiib-rium constants are the heats of evaporation of methanol, iso-butylene, and MTBE, at the standard state, as well as theirmolar heat capacities in the formof power functions. TableI shows ailthese thermochemical data. The required standardenthalpy and entropy changes of the reactlon inthe gasphasehave been taken fromour previous work (6): A H o @ )=-65.4kJ mol-' and AS', =-174 J mol-' K-'. Thermochemical datayield the following values at 298 K by usingeqs 4-7: AH' =-37.1 kJ moi-l, AGO =-13.5 kJ mol-l, and AS' =-79.3 Jmol-' K-l. From the first two, an equation for the theoreticaltemperatwe dependenceof the equilibrium constant, usingeqs10 and 13, is obtained:in K = 1145 - 14740T-'- 233 in T + 1.066T-1.077 X 103T2 i- 5.306 X 10-'T3 (14)The values of the theoretlcal equilibrium constant predicted byeq 14 are shown In Figure 3 (dashed line).

7/27/2019 julito4

3/5

Jormel of Chemical and Enghewhg Data, Vd . 37,No. 3, lQQ2 34 1

5-

~0.0280.0320.0230.0070.0630.0370.0430.0330.0710.0490.0220.0150.0140.0830.0610.0310.0210.0200.0760.0420.029

Table 11. Experimental Conditionsan d Obtained Equilibrium ConstantsT/K bA/l XM XI XE K, YM Yr YE K.1 K

1 313.7 1.06 0.076 0.896 426.3 2.773 1.265 1.006 0.287 122.323456789101112131415161718192021

313.7313.7313.7323.5323.5323.5323.5333.3333.3333.3333.3333.3343.2343.2343.2343.2343.2353.2353.2353.2

1.23"1.45"1.501.001.061.23"1.45O1.001.061.271.481.501.001.061.271.481.501.061.271.48

0.0720.0950.2950.0580.0930.0900.1130.0810.1120.2540.3420.3560.1060.1340.2750.3670.3830.1600.2960.389

0.8950.8820.6970.8790.8700.8680.8540.8470.8390.7240.6430.6300.8110.8050.6940.6120.5970.7640.6620.582a MTBE iswed a t the beginning.

Rosuttr and DkcusslonThe experiments were carried out at t he temperature range313-353 K, under a pressure of 16 bar and with an intlalt?"ol/fsobutene ratio from 1 to 1.5. Thatmolar ratio rangewas chosen seeing that, for methanol/isobutylene ratios lessthan 1, the byproduct dlisobutene was significantly produced,even atbw temperatues,andwhenthe ratio was greater than1.5 the undesirable dimethyl ether was also formed.The mass balance was checked by comparingthe molarvariations of isobutene, methanol, and MTBE. The averagediscrepancy was about 5%, which couM be explained byconsidering the small losses (by evaporation) of specles onm t o h ~he pressureafter t hechemical analysis, andthe smallamount of isobutylene changedintodilsobutylene that probablywas inthe detection boundery of the chromatograph. We as-sumedthat the nitrogen amount dissolved in the reactant mix-

hre(about .6%) was ne&ibk anddid not affect t hechemicalequilibrium at aU.Table I1 shows the experimental conditions and results aswell as the nonkieai behavlor of t he system calculated by theUNIFAC methad. It Is computed that the experimentalerrorof mdar fractions is f0.002,which allows the expression oft hemolarfractions to three decimal places. The amounts ofdiisobutene and dimethyl ether detected at the highest tem-peratures, only inthe runs inwhich tsobutene or methanolwereused in great excess, were in the boundary of the chromato-graphmass detwtbn. Those quantltles yle#edmolar fractionsless than or equal to the experimentalone mentioned above,particularly diisobutene whose molecular weight is high.Moreover, their effectonthe actMty coefficients of methanol,isobutene, and MTBE was negligible. Hence, the dlisobuteneand dimethyl ether amounts were not presented in Table 11.As can be seen, MTBE activity coefficients are very closeto unlty. In umtrast, isobutylene and, malnly, methanol activitycoetfidentsare higher than 1. Thevalidity of UNIFAC predic-tlons for that reacting system was already checked (2). Theparameters needed for the use of UNIFAC have been takenfrom the tables published by SkjoWorgensen et ai. ( 9 ) ,QmhYng et ai. (IO), M t ai. (77) , and Tiegs et ai. (72).It is worth nothing that at each temperature theKvaiuesob-tained were not affected by the InWi molar ratio, under 0th-mlse uniformconditions. Also, these pseudoexperimentalequilibrium constants followed the general trend of loweringwhen temperature increased, as expected for an exothermicreaction. Therefore, we can conclude that the experimental

382.3 2.813 1.258 1.005 0.284 108.6397.9 2.642 1.292 1.010 0.296 117.7319.8242.7250.4227.5226.9146.3153.7128.4126.8125.892.898.081.979.779.762.753.651.5

1.7462.9382.6342.6672.5102.7332.5041.8651.6161.5842.5602.3741.7951.5601.5242.2461.7381.515

1.6801.2211.2751.2691.3041.2401.2881.5421.7491.7851.2611.3051.5651.7871.8341.3281.5861.818

1.1121.0011.0081.0071.0131.0031.0111.0741.1391.1511.0061.0141.0821.1531.1691.0191.0891.165

0.3790.2790.3000.2980.3090.2960.3130.3730.4030.4070.3120.3270.3850.4140.4190.3410.3950.423

121.267.875.267.770.243.348.248.051.151.228.932.131.533.033.421.421.221.8

r=1.27fc40 0.93-

r=1.48

- 0.89

k I I +

0.8640 50 60 70 80 1T i0CFbwo 2. EqulYbrlum isobutene conversionsatdifferent temperatures.

- 1

22.8 2.9 3 3.1 3.2103/(/K) 3Figm 3. h Kagalnst1 / T . Comparison betweenthe values obtalnedexperimentally (solid line) and those predicted from thermal data(dashed line).method followed and UNIFAC predictions are suitable forequitbiumconstant determinations, at leest, forMTBE syntheskor decomposltkn. A comperlson, owever, wlth theoretical andliterature data is necessary to back up that conclusion.Figwe 2 is a plot of the lsobutene equilibrium conversion asa function of 1/T. A maximum conversion is obtained at 40.5OC for a methanoi/isobutene ratio of 1.5.

7/27/2019 julito4

4/5

54 2 Jownal of Chemicaland Englneerlng &fa, Vol. 37,No. 3, 19921 1 0 4 , I

llm!

2.8 2.9 3 3.1 3.2 3.3l d I ( T I K )10%

Flgun 4. InK+f (T) against 11T. Comparison betweenthevaluesobtained experimentally(solid line)and thosepredictedfrom thermaldata (dashed line).If the entheipy change of reactkmis assumed tobe constantover the temperature range, the temperature dependence ofthe pseudoequilibrium constant can be found by using theequation

in K =A S o / R - ( A H o / R ) ( l / T )which Is deduced fromeqs 3 and 4. Thus, when the Kvaluesin Table I 1 are fitted to this equation, we can calculate AH ofrom the slope and A S o from the intercept. Obtained valuesare

AH o =-39100 f 800 J mol-'A S o =-85.3 f .5 J mol-' K-'

The solid line in Figure3 Is the line fitted to pseudoexperi-mental data, and the dashed line denotes the InKvalues pradlcted theoretically by eq 15 with AH o =-37.1 kJ mol-' andA S o =-79.3 J mol-' K-' . A very slight difference may beseen betweenthe two lineswhich lmpiles that the experimentalslope Is slightly greater than the theoretical one.If , however,the variation of the enthalpy change of reactionis significant, eq 14 mustbe usedto fit the data (Figure4). Inthis case, we obtain I , from the slope and I , from the lnter-cept, and therefore, a more correct expression for the tem-perature dependence of K isInK =1144 - 14634T-' - 233 In T + 1.066T-1.077 X 10-3T2+5.306 X 10-'T3 (16)In Figure4, Itcanbe seen thatthe difference between the solidline (experimental values) and the dashed line (theoretical val-ues)Is smaller than in Figure 2. This fact indicates that AHovariation with temperature Is enough to take into account onworking out equilibrium constants. Rehflnger and Hoffman ( 7 )foundtheoretically that AH o Is In the range -37.7 to -43.6 kJmol-' for 298 and 363 K, respecttvely. At the same tempera-ture range we found for AH o -37.1 to -43.1 kJ mol-'.From eqs 3, 4, 10, and 13, equations for the temperaturedependence of A H o , A S o , and AGO can be obtained:

1.792 X 10-2T3+1.3230 X 10-5T4A H o (T )= 121670 - 1935T +8.8598T2 - J mol-' (17)A S o ( T )=7583 - 1935 In T + 17.72T-2.687 X 10-2T2+1.765 X 1 0 - v 3 J mol-' K-' (18)A Go (T ) = 121670 - 9518T+ 1935TIn T - 8.8598T2 +8.958 x 10-3T3- 4.4113 X 10-9' J mol-' (19)

Table 111. Standard Free Energy, Enthalpy, and EntropyChanges of MTBE Synthesis in the Liquid Phase at 298 KAHO(298 K)/ ASO(298K)/ AG0(298K)/(kJ mol-') (J mol-' K-') (kJ mol-')

AHo function of T -38.0 f 0.8 -81.7 f 0.5 -13.6 f 0.8theoryAHo constant -39.1 f 0.8 -85.3 f 0.5 -13.7 f 0.8theory (this work) -37.1 -79.3 -13.5

Rehfinger ( 1 ) -37.7 -14.0Arntz (13) -39.2 f 0.4Gicquel ( 3 ) -39.8 f 2Gupta (15) -36.8Obenaus (14) -37.0

Table IV . Obtained Equilibrium Constants and ThosePublished by Other ResearchersRehfinger Al-Jaral-T/"C present work ( 1 ) Colombo (2) lah ( 4 )

40 l l 8 f 9 136 15150 70 f 5 85 10060 48 f 4 54 6870 32 f 2 35 47 3880 21 d2 1 23 33 1690 16 24 13100 11 18 7Table 111shows the values of AGO, AH', and A S o at 298K determined for AH o assumed to be constant and for AH ovariablewtth the temperature, along wlth those theoretical andexperimental values deduced in other works (7, 3, 73-15)using different methodologies and resins. As can be seen, allthe values of A Ho are quite similar at 298 K, within the limitsof experimentalerror, WMChaccounts, again, for the validity ofthe methodology used In this work. Nevertheless, If values ofKare calculated at different temperatures (TableHI), e cansee discrepancies that cannot be explained unless a AH ovarlatbn with temperatureIs assumed(7). It may be seen InTable 111thatthe values of Colombo et ai. (2)are greater dueto the fact that theyusedan almost constant AHo of -34.7 kJmol-'. The theoretical equilibrium constants calculated byRehfinger and Hoffman are slightly different from those deter-

mined by us experimentally,butthesediscrepancies come outon changing the temperature due to, likely, small differentvalues of AHo. Finally, the Kvalues determined by Ahlarailahet ai. (4)are also qulte similar to ours, and the discrepanciescan be explained If we take Into account that they covered ahigh temperature range where we found problems in getting atrue equillbtium owing to the presence of side reactions at suchtemperatures.ConclUrlOIlSThe agreement of AHo obtained experimentally In this workat 29 8 K with those determined in other works using differentmethods In some cases proves that the methodology used iscorrect. The trends of the Kvariations with temperature con-firm this fact.From the temperature dependence relationship for theequllikkrm constant, equations for the temperaturedependenceof enthalpy, entropy, and free energy changes of the MTBEsynthesis have been determined. These values show that thevariation of reactkm standard enthalpy with the temperature isimportant In equilibrium calculations.Theexperimental equiliklum constants at lower temperature(40 and 50 O C ) determined in this work cover a small gapexisting In the open literature for MTBE synthesis In the liquidphase.

Notatlona , b , c , d = changes of molar heat capacity coefficients wehchemical reaction

7/27/2019 julito4

5/5

J. Chem. Eng. Dei% 1002, 37, 343-349 343e,,b,,c,,d,=coefficients intheequation for molar heat capacity v =vaporization. . .of component i ; C, =a, +b, i+c, T' +d T 3=molar heat capacity of component i , J mol-' K-l3 =standard free energy, J mol-'H o =standard enthalpy, J mol-'I , =Integration constant in van't Hoff's equation, adlmensionaiI , =integration constant in Kirchoff's equation, J mol-'K =thermodynamc equilibrium constantK , =equlllbrium constant based on activity coefficientsK, =equilibrium constant based on molar fractionsMTBE =methyl fert-butyl etherr =molar ratioR =gas constant, J mol-' K-'So =standard entropy, J mol-' K-'T =temperature, Kx, =molar fraction of componenti&e& Symbolsy,=acthrity coefficient of component iPOo =standard free energy change of reaction, J mol-'AHo =standard enthalpy change of reaction, J mol-'AH',,A S o =standard entropy change of reaction, J mol-' K-'Y, =stoichiometric coefficient of component iSubscriptsE =MTBEe =equilibriumg =vapor phaseI =isobutenei =componentI =liquid phaseM =methanol

standard heatof vaporization of component i , J mol-'

~ l g# l yO. m, 634-04-4; K-2631, 1206 69-569 ; WH, 6 7 - w i ;isobutene, 115-1 1-7.Literature Cited

(1) Rehflnger, A.; Hoffmann, U. Chem. Eng. Sci . 1000, 4 5 , 1808-17.(2) Colombo, F.; Corl, L.; Dalloro. L.; Delogu, P. I n d . Eng. Chem. Fun-darn. 1083, 22 , 219-23.(4) AMarallah, A. M.; Slddiqul, M. A. 6. ; Lee, A. K. K. Can. J . Chem.Eng. 1088, 66 , 802-7.(5) Marsman, J. H.; Panneman, H. J.; Beenackers, A. A . C. M. Chsomefo-graphia 1088, 26 , 383-6.(8) Tejero, J .; Cunlll, F.; Irqulerdo,J. F. I nd . Eng. Chem. Res. 1088, 27 ,(7) Reld, R. C.; Prausnitz, J . M.; Sherwood,1.K. The PrOperHes ofQmes( 8 ) Denblgh, K. 0. Wnclpks of ChemrCelEquHlbrhKn,4th ed.; Cambridge(9) SkJolbJorgensen, S.; Kolbe, 6. ; Qmehllng, J .; Rasmussen, P. Ind.

( I O ) Gmehllng, J.; Rasmussen, P.; Fredeslung, A. I nd . fng. Chem. Roc.(11) Almelda, E.; Weldllch, V.; Gmehllng, J.; Rasmussen, P. Ind. Eng.(12) Tlegs, D.; (knehllng, J.; Rasmussen. P.; Fredeslung, A. Ind. E r g .(13) Arntz. H.; QotUieb, K. J . Chem. 77wm@yn. 1085, 77, 967-72.(14) Obenaus, F. f rdoelohb E M S , etrochem. 1080, 3 3 , 271-75.(15) Gupta, J. C.; hakash, J . Chem. Eng. World 1080, 15, 27 .(18) Gallant, R. W. phvslcel properties of & d m a r b o n s ; Qulf Publlshlng(17) Rowllnson.L/quhfs andL/quhfmixtwes,2nd ed.; Butterworth: London,

Received for review Decemb er 13, 1991. Accepted March 18 , 1992. Weexpress thanks fo r flnanclal support of part of this work to the refining com-pany REPSOL PETROLE0 S.A.

(3) GlCqUel, A.; TWCk, 6 . J . Ca tel . 1083, 83 , 9-18.

338-43.andLiquMs, 4th ed.; McQraw-HIII: New York, 1987.Unhrerstty: Cambridge, U.K., 1981; Chapter 4.Eng. Roc.Des. D e v . 1070, 18, 714-22.Des. D e v . 1082, 21, 118-27.Chem. ROC.es . Dev . 1083, 22. 878-81.Chem. Res. 1087, 26 , 159-81.

Co.: Houston, 1970.1980.

Hydrates of Hydro carbon Gases Containing Carbon DioxideSanggono AdlsasmHot and E. Dendy Sloan, Jr."ChemicalEngineeringand Petroleum Refining, Colorado SchoolofMines, Golden, Colorado 80401Thk work presents equlllbrlum measurement8 for hydratethree-phase (vapor-hydraleaquoour Ilquld) behavior fornatural gas componentswith hlgh concentratlons ofcarbon dk xld e, at 1.mp.ratur.r between the Ic e polnt andih e quadruple point of each m Mu re. Blnary hydratepha8m oqullb rla were measured for carbon dloxlde witheach of the fdk wln g g am e methane, ethane, propane,ko bu tane, and normal butane. Data fo r three of thoseblnary 8yrtema (carbon dloxlde +ethane,2meihylpro pane (kobu tane), and butane) do not exb i Inthe literature. Our resuits for the other two blnarysy r to m are shown to complement or supplant parilal datau t a rom other Iaboratorles. I n additlon, resutts arepresented for synthalc muitlcomponent mlxturescontalnlng flxed ratlos of methane, ethane, propane,2.thy lp ro pan e, and butane, with varylng amounts ofcarbon dkxld e. A molecular Interpreiatlon k given Interms of hydrate rtructural properties.Introductlon

Carbondkxlde hydratee have several unique applications. Inthe earth's permafrost and deep oceans, there is a large re-source depositedas hydrocarbon and carbon dioxide hydrates.To whom correspondence should be a d d r e d .

gy Bandung. Bandung 40132, Indonesla.Present address: Chem ical Englneetlng Department, Instl tute of Tech no b

Kvenvolden( 1 ) n 1988 estimated that the amount of carbonin hydrocarbon hydrates is on t heorderof 10000Qtons (1 Qton= 1 X 1015g), greatly surpassing the resource of ail othercombustible fossil fuels. Since carbon dioxide is normallyproduced naturally with hydrocarbon, hydrate phase equiiibrlumdata are necessary to determine both the extent and futurerecovery of this energy resource. Recently carbon dioxideandhydrocarbon outgassing from such in situ hydrates has beenhypothesized to contribute to the global warming process (7).In addition to such natural processes, carbon dioxide playsa substantial industrial role in gas production and processing.Carbon dioxide is used in enhanced oil recovery (EOR) pro-cesses, and processing the associated gas involves a risk ofhydrate formation. Many natural gas wells produce gas withhigh carbon dioxide concentrations, such as the wells at theLaBarge reservoir in western Wyoming, andthe Natuna pro-duction field in Indonesia.Evenwith these needs, only two binary systems containingcarbon dioxide have been reported in the literature, namely,carbon dioxide+methane, measured by Unruh and Katz (2)in 1949 and Berecz and Baila-Achs(3)n 1983,and carbondioxide+propane, measured by Robinson and Mehta (4) in197 . Measurement details of binary mixtures of C 02+CHwere shown to confirm and extendthe Unruh and Katz data inour previous paper (5).No publlshed data exist for binary mixtures of carbon dioxidewith etther ethane, P-methylpropane (isobutane), or butane. Thelatter two binaries and the COP+C3H, binary form bothstructure I nd structure I 1 hydrates at high carbon dioxide

OQ21-9568/92/1737-0343$03.QOfO 0 1992 Am er i can Chemic a l Soc iety