CA2174347A1 - Sour gas treatment process - Google Patents

Abstract

Improved membranes and improved membrane processes for treating gas streams containing hydrogen sulfide, carbon dioxide, water vapor and methane, particularly natural gas streams. The processes rely on the availability of two membrane types, one of which has a hydrogen sulfide/methane selectivity of at least about 40 when measured with multicomponent gas mixtures at high pressure. Based on the different permeation properties of the two membrane types, optimized separation processes can be designed.

Description

~7~3~7 ~0 gS/11738 PCTIUS94/12099 SOUR GAS TREATMENT PROCESS

FIELD OF THE INVENTION
The Ul~ c .Ition relates to ~"occSScS for l~.llVVIllg acid gases from gas streams. More pal L;-,ul&.ly, S the i.l~ltioll relates to a In~ la IC process for ~...uv-ng L~nLu~,.l sulfide and earbon dioxide from gas streams, such as natural gas.

BACKGROUND OF THE INVENTION
Natural gas l.r~vidcs more than one-fifth of all the primary energy used in the United States.
10 Much raw gas is "s~lbq~ ity", that is, it exeeeds the pipeline sl~ ;l;~ ~;o,~c in nitrogen, earbon dioxide and/or hydrogen sulfide eontent.
The best L-~,dt~ L for natural gas right now is no ll~àl~llenL. Raw gas that is known to be high in nitrogen content, high in nitrogen plus earbon dioxide content, or high in L~J1U~ I sulfide content is usuallylcflintheground,becauseitcannotbee,.lla~t~landtreatedc~..o.~ llywithpresent~lu~s,;llg ter~ olo~.
There are several aspects to the problem of treating natural gas to bring it to pipeline specifications. The first is the removal of imyul;tics~ primarily water, l.~d.ugcn sulfide and earbon dioxide; the seeond is loss of methane during ~IUC~ g P~.,ccssc.s that remove h~c n sulfide and earbon dioxide may âlsO remove a portion of the methane. Losses of less than about 3% are normally acceptable; losses of 3-10% may be a: :c.~ e if offset by other a.lv~lta~ " losses above 10% are nu"~lally ~ r' 1 1~ A third aspect is the fate of the illl~ iCs onee r~,-llo-cd. Carbon dioxide can be ~Lal god or . ~ ; d, but hydrogen sulfide, which is toxic even in low COIl~l~l dtiùl~ " must be treated.
If thc waste strea n C4~ g II~U~ I sulfide can be Cull~ lll ~ Ylfficiently, it may be passed to a Claus plant for cu"~ ;on to sulfur. Waste streams co~ Iow con~dl~4tions must be .~ Jn~. ~ of in some other way, sueh as a redox process of the LO CAT or Stretford type, for , 'e, or, less desirably, flaring.
Choice of ~p,u~ .t~ hl- ,.lt is, therefore, not straightforward, and depends on the feed gas coln~,osition, the size and location of the plant and other variables.
When natural gas is treated, most plants handling large volumes of sour gas co~ g greater than about 200 ppm hydrogen sulfide use amine-based h ~' ~ Ic~ for acid gas removal. Amines co.. ~ol~ly used include MEA, DEA, DIPA, WA and MDEA. The plants ean remove both earbon dioxide and h~u~ l sulfide. When the amine solution is spent, the acid gases are flashed off and the WOg5/11738 217~47 Pcr/uss4/12099 0 solution is ~ ,...t~l The - ' ' c.~ .... .1 in an amine plant makes it ausc~ ible to failure. The plant includes heaters, aerial coolers, pumps, etc. and requires frequent quality checks and ~ t~ G, making opcrational reliability probably the weakest feature of the le ' ' ~
Amine plants do not sorb methane to any ~ l extent, so methane loss is not an issue in this 5 case. However, the LyJIu~w--sulfide- in$ gas stream Inudu~;l when the sorbent is l~ al~ must still be treated, subject to the same COI al,a;llts as above.
As an alternative to amine sorption, or as a po~ step following any process, ~poçi~li7 SCa~_~l~lg or sulfur ,~cu.~,.y plu~,C;~aCS, such as Sulfa-Scrub, Sulfa-Check, Ch~ .~l, Supertron 600, solid iron sponge or solid _inc oxide may be used for low-volume streams c .. 1,~ less than about 10 100 ppm hydrogen sulfide. Many sca~...gc.a present s~ l disposal ~ bl~,...s, however. In an i..~,.~,aa;..g number of states, the spent sca~,.l~. c~ t~ ~ toxic waste.
A culls,d-,-able body of literature exists ,.,g~u.l;..g ,..e.-.~- - based 1-~ d~ lt of natural gas, mostly using ce~ lose acetate (CA) ~ CS to remove carbon dioxide. Although cellulose acetate Ill~nll.lall~, plants are designed to remove carbon dioxide, cellulose acetate Ill~,.llI)lallcs also have scl~liv;l~
IS for hydrogen sulfide ova mcths~le~ so they tend to cc~ a~L small amounts of hydrogen sulfide. Unless the raw gas stream contains very high col~ .t;olls of carbon dioxide, however, it is not possible to reduce a stream C4~ g even modest amounts of hydrogen sulfide to pipeline 5y~ ;ol~ (usually 4 ppm hydrogen sulfide) without vastly O.~ /lu~,C;~a~lg as far as the carbon dioxide sp~ifi~ti~n is concc...cd. If such U~ JIUccaa;llg is perrormed, large amounts of me~hane are lost in thc ...~,...1,.
20 p.,.lll.,àl~ stream, and this is normally ~ a r: b'c Only a few of the many literature ~,L~s relating to llh,-l~l - based carbon dioxide Ll t.~dt~ n, -r r lI,y discuss ~moval of hydrogen sulfide in c~ with the carbon dioxide. A paper by W.J.
Schcll et al. ("Scy~ of CO2 from Mixtures by Me..~ P. s ,;.c..l~ at the Gas Cor ' ~r~Cu~lfww~, UlU~ail~ of O~ ms Marchl983)sâysthat"IftheH2Slevelislowenough, 25 the IIIC.II~ IC system can âlsO be used to meet pipeline specification for this c~ without any furthcr ll~idt-ll~,.ll required." The paper shows a case where a c " ' : acetate ...~,...1" system can be used to reach pipeline spe~ ir - for carbon dioxide and h,l.~gcn sulfide in two stages, starting with a feed content of 15% carbon dioxide and 250 ppm Lydl~gen sulfide, and points out that, for high cu"~.~l~..tions of L.~Lûgen sulfide, "a much larger number of ~ c are required to reduce the H2S
30 levels to pipeline ~y ~ , (1/4 grain) than for C2 (3%)." The costs of membrane tl~,dtnl~,.lt are c~ ,d to be more than 100% higher than co~ diondl amine ll~,dtl...,.d in this case.
A report by NN. Li et al. to the Department of Energy ("Mc.. ,l., - S~ ,s in the wo 95/11738 2 1 7 4 3 ~ 7 PCT/us94/l2099 Pctrocke ~ r~l Industry", Phase II Final Report, S~t~ l.c~ 1987) e~ the effect of illlyuliliC5, ; ,g hydrogen sulfide, on the ability of cp-~ nce acetate lll~ l ..,s to remove carbon dioxide from natural gas. The ~,yu-t~ i found that the membrane perf~i~sirire was not affected significantly by hydrogen sulfide alone. II~ , dramatic loss of ~,~ "l - y- - ---~ nbility was ol~s.,. ~,cd if both Ly~Lug~
5 sulfide and water vapor were present in the feed. The authors co-~ -A~A that "s~ r. ~l use of these CA-based --- ,~ r.C must avoid y-u~---g gas which simult~ly has high H20 and H2S concentrations".
Anotherproblem s ~ Y -d with c~l' ' ~ acetate l..~ is water, which is slways present in raw natural gas streams to some extent, as vapor, C~ à~ liquid, or both. The gas separation properties of cell~lose acetate membranes are dcaLI~J~l by contact wi~h liquid water, so it is normally 10 noec;,:,a.y to provide yl~ dt ..~ to knock out any liquid water and to reduce the relative humidity low enough that tlhere is no risk of co~ ;o.~ of water within the membrane modules on the yl - side.
Fori , 'c, the above-citedpaperbyW.J. Schell et al. ("S~ of CO2 from Mixtures by Membrane pl ....~ ~;o..", y,~lt~ at the Gas Conditioning Ccll~w~, U~ of Ol~ n,- 7~, March 1983) points out that "Even though Ill~,llltila lC systems ~ -ly dehydrate while ~,---ov-..g CO2, care must be 15 taken to avoid ~--1~ g the ~ .anc with liquid water. Feed gas streams S~ lldt~ with water are normally p.~ at~ d to at least 10 above the water dew point at the feed inlet y .,~ and the pressure tubes and inlet piping are insulated to prevent condensation."
The above-cited report byN. N. Li et al. (nMe..lk. - Se~ dtion P~v_c;,~.s in the r~ k....~
Industry," Phase II Final Report, Sc~ t ---~ 1987) presents data showing the effect of water vapor on 20 lll~,l.ll,l.ul~ flux for cellvlose acehte ...~...b. a les, and c~ ~h,-~ that "for relative kv- ..;~li1 ;-,s of 30% and higher, the flux decline is large, rapid, and i--~ . I.le". E.W. Funk et al. (HEffect of T..,l.... ;fi. C on Ccllnlosc Acetate Mc...l).- ~c Perfnnnnnee", Recent Advances in Sr.~. ~1;n., Teehniqlles - III, AIChE
S~ .. Scrics,250, Vol 82, 1986) advocate that "Moi;,lu~ levds up to 20% RH appear tr'~ dl~lt but higher levels can cause i...,~ le ..e..-b-an~, compaction".
U.S. Patent 4,130,403 to T.E. Cooley et al. (Removal of H2S and/or CO2 from a Light H~ ~.StreambyUseofGas Pc .~'-le Membrane, 1978,Col. 12,1ines36-39)statesthat"Ithas been d;~u . _~l that in order to function errt~ _ly, the feed gas to the c~ ,ose ester membrane should be s.ll,s~ lly water froe". A second paper by W.J. Schell et al. ("Spiral-Wound r.-...~ n.~ for plmfir~tinn and Recovery", Chemical F-~g;..- - . ;..g Progress, October 1982, pages 33-37) u~ ....c that 30 "Liquid water is d~,l-i..._nlal to the performance of the membrane, however, so that the feed gas is dd;~ l to the ~ ,lllbl - system at less than 90% relative hu.. d;ty."
ln other words, ~l~hnugh oell~llose acetate membranes will p~ ~ water p.~,f~ 1idlly over WO95/11738 Pcrluss4ll2o99 mf thslne~ and hence have the --r 1 lity to dchy~L the gas stream, care must be taken to Iceep the amounts of water vapor being ~ -~ low, and, a~.g to some If 3~ g~, as low as 20-30% relative humidity.
In light of these l;. . .; ~ , COllS;d~ effort has been e ~ fl~ over the last few years in the S search for ..- -..l.... ~ materials that would be better able to handle streams u...~ g carbon dioxide plus sflulldal~ cv..l~ ; notably hydrogen sulfide snd wster.
For dense polymer ~ , the ~ b:-~r~l cffect of the sorption gnd ~ h ~
d~l._....;-..,s the sel~Liv;l~ of the l.~,..lLl - The balance between mobility, or rliffilci~n, sel~livll~ snd sorption self~liv;l~ is different for glsssy snd rubbcry pOl~..l.,.a. In glassy polymers, the mobility term 10 is usually ~lo...;~-- .l, p., ~ bility falls with i,l~,..,aai..g pn-,... .I size and small ^'e ~es p.,.ll,e..le preferentially. In rubbery polymers, the sorption term is usually dc....;..~-.l p~,.lll~,ability ill.,.~ases with ill~,l~aaillg p~.ll~d size and larger -~- nlle5 p~ preferentially. Since both carbon dioxide (3.3 A) and hydrogen sulfide (3.6 A) have smaller kinetic ~ . than methane (3.8 A), and since both carbon dioxide and hydrogen sulfide are more cr~ lc than mf th~n~, both glassy and rubbery r ~l -~
lS areselectivefortheacidgasc~ l;over tl - Todate,however,mostllle.lll,lalicd-,~ lo~
work in this area has focused on glassy materials, of which ce~ sse acetate is the most ave~fi~
r lC
In citing ~l~liv;ly, it is i...~ to be clear as to how the pi ~ en data being used have been III~Ul~l. It is cv .~ .. to measurc the fluxes of difi^crent gases a~,lJa. l~/, then to c~'~ ' s~ ivi~/
20 as lhe ratio of the pure gas p - ,,,..~ This gives the "ideal" scl~livi~y for that pair of gases. Pure gas ll.caaul,.ll~,.lla are more cr~ ly reported than mixed gas c,.~.ill.~,.lla, because pure gas eA~efilll~da are much easier to perform. ~ g the ~-~-s~;o~ data using gas llli~ ,s, then c~lr~ ing the sele~,livily as the ratio of the gas fluxes, gives the actual sclc~,livily that can be achieved under real conditions. In gas mixt~cs that contain ~ v. ~ ;, it is ~u~llly"~l~hmt~h not 25 always, the case t-h-at the mixed gas sel~liv;l~ is lower, and at tirnes colla;de.ably lower, than the ideal ~I~Livhy. The ~ u,- ~, which is readily sorbed into the polymer matrix, swells or, in the case of a glassy polymer, p~ t;-, ~r~ the membrane, thereby ~I~u.g its Lsc.l.lu lating cap ~ itiP~
A t~,~,h~ c for pl~li~,Li..g mixed gas ~ f~ -: under real a~ntlitirm~ from pure gas lllCaaul~llwlb with any ~,I;~iLty has not yd been d~ ,lu~d. In the case of gas I~uAhu~ sutch as carbon 30 dioxide/meth~nP, with other ~v~ ;, the e --r _' '- iS that the carbon dioxide at lesst will have a swelling or p~ ;-.g e~ffect, thereby ~~, I,G the membrane ~ l..- ~ cI.~a~ ;sL~,s. This e~l ~l,. ion is borne out by c~ se acetate membranes. For example, 3CCOIding to a paper by M.D.

~Wo 95/11738 2 ~ 7 ~ ~ 4 7 PCrlUS94/12099 S

Donahue et al. ("P~....~ ';on ~hav;Or Or carbon l;o~de methane llli~t~U~,a in c~ r- acetate membranes", Journal of Membrane Science, 42, 197-214, 1989) when ll~aaul~,d with pure gases, the carbon dioxide ~ y of ~ .Ll;c c~ ose acetate is 9.8 x 10-5 cm3/cm2 s kPa and the methane ,ve~ h;l;~ is 2.0 x 10~ Wll~ -S LPa, giving an ideal sel~liv;ly of about 50. Yet the actual sel~Lvily S obtained with mixed gases is typically in the range 10-20, a factor of 3-5 times lower than the ideal sclccliviLy. For eY~np~e~ the report to DOE by Norman Li et al., l;C-,~e~ above, gives carbon Lhanea~ iviticsintherange9-15forOnesetOffieldtrialS(at6,000-6,240Wa(870-905psi)feed pressure) and 12 for another set (at 1,483 Wa (200 psig) feed pressure) with a highly acid feed gas.
The W.J. Schell et al. Ch~m;rs~l F~ Progress paper, ~ e~ above, gives carbon dioxide/mcthane s~,l~liviti.,s of 21 (at 1,828-3,207 kPa (250-450 psig) feed ~ aulc) and 23 (at 5,620 kPa (800 psig) feed pressure). Thus, even in mixed gas ll.~UI~ t~, a wide spread of scl~liv;lie5 is o~t~;n~, the spread .1~ u~; g partly on ~d~llg c~n~litit n~ In particular, the pl9 ~~ g or swelling effect ofthe carbon dioxide on the membrane tends to showpressure ~ alth~ h it is sc--.- l; ~ s hardto~ hthisfromothereffects,suchasthecoltlib-ltio-ofsee~n(' yc~ ' m~lc c~ or~
The search for illll,lo. d membranes for l~ .illg acid com~,o~ t~ from gas streams, n~
it has focused primarily on glassy membranes, e~r4~ 9~ several types of mernbranes and membrane materials. A paper by A. n~Sfh~ pS et al. ("D,. ~ P-~I of Gaseous P~ ;c.n Membranes adapted to the Purification of Hydrocarbons", I.I.F - I.I.R - C.. ~ o.. A3, Paris, 1989) ~3esG~ibes work with u~atiC polyimides having an intrinsic material selc~ Livit~/ of 80 for carbon dioxide over methane and 20 200,000 forwatervaporover~h-s~n~ Thepaperdefines the target sel~liv;li~s that the l~,scal~.llw:~ were aiming for as 50 for carbon d;u ddc~ ' ~ and 200 for water vapor/ -- ~h~-~e The paper, w_ich is principally directed to dehydration, docs not give carbon ' * /~ llallc selccLv;li~,s, except to say that thcy wc~ "gcnally low", even though the .,.~ il...,.lt~ were carried out wilh pure gas samples. In other words, despite the high intrinsic sclc~livily of 80, the lower target value of 50 could not bc reached.
British Patent number 1,478,083, to Klass and T ' ' 1, presents a large body of p~ n data obtaincd wilh ~e~ c/c~l,ol- diu~ idc~dlu~ l sulfide mixed gas streams and polyamide (nylon 6 and nylon 6/6), polyv~yl alcohol (PVA), poly~lu litrile (PAN) and gelatin -- -.1,. ~ Some ~ ,t~Aly high sd~livili~,s are showrL For the nylon --- --,l--.. Irs carbon dio~id~ ' ~ sel~LviL.,s of up to 30, and hydrogen sulfide/methanc selc~,Lv;L~ up to 60, are reported. The best carbon ' scl~livily is 160, for PAN at a tcmperature of 30C and a feed p~ ; of 448 kPa (65 psia); the best l~y~ ,sulfide/methane selc~livily is 200, for gelatin at the sa ne cQ~ ;o.~c In both cases, however, the p~,... ~hility is e.~ .llcly low: for carbon dioxide through PAN, less than S x 10~ Barrer and for W O 95/11738 2 ~ 7 ~ 3 ~ 7 PCTrUS94/12099 ~

hydrogen sulfide through gelatin, less than 3 x 103 Barrer. These low p~,.,-,eabilities would make the llal~C~ c fluxes ,.-.s., ..~'e for any practical I~UI~USCS. It is also ~ .., whether the gelatin Ill~n~lanc~ which was p~ 1 with glycerin, would be stable much above the modest pl~,aaUl~,a under which it was tested.
U.S. Patent 4,561,864, also to Klass and T ~ ;IlC~ uldteS in its text some of the data reported in the British patent .1;~ ~ above. The ~864 patent also includes a table of c ~ ,.c for ce~ lose acetate ",~ . , showing the rel ', between "Figure of Merit", a quantily used to express the purity and methane r~cu~ y in the residue stream, as a function of "Flow Rate Factor", a quantity that appears to be ~ hdl akin to stage-eut. In plrf} g the ealculations, a~/al~liOI) factors (where ~he sc~ io" factor is lhe sum of the carbon d;~"~idc/~ e sclc~Livily and the hydrogen sulfide/melhane scl~livily) or20 to 120 are ~CcllmP~ The figures used in the cS~Ic~ tinnc appear to range from the low end of the c~"nb;"ed carbon dioxide and hydrogen sulfide scle~livili~s from mixed gas data to the high end of the .,on,l.u,cd scl~liv;li. s ~ from pure gas data.
A paper by D.L. Ellig et al. (IlCon~.~L dtiO~. of Mcll. u.e from Mixtures wvith Carbon Dioxide by ~c---,ca~ion through Polymeric Films", Journal of Me.. ~. - Science, 6,259-263,1980) alUIlnlal;~a .. . tests c~arried out with 12 different w.".-,.,.-,;ally available films and ...~,..11,. .u~es, using a mixed gas feed ~ _ 60% carbon dioxide, 40/O methane, but no h~Lug~ sulfide or water vapor. The tests were carried out at 2,068 kPa (about 300 psi) feed pressure. The results show a~,lc~,liv;lies of about 9-27 for eelllllose acetate, up to 40 for pol~ aulfonc and 20-30 for polysulfone. One of the membranes 20 tested was nylon, which, in ~ L ' to the results reported by Klass and T.~nf~:~hl, showed ecc~nti~lly no sel~liv,l~ at all for earbon dioxide over m~th~ne~
The already much~ic~ DOE Final Report by NN. Li et al. eontains a seetion in which separation of polar gases from r,o.. F " gases by means of a mixed-matrix, faeilitated halls~Ju,l men~ IC iS ~ 1 The membrane eonsists of a silicone rubber matrix earrying pol.~ l.,nc glyeol, 25 which is used to f~ilits-t~ transport of polar gases, sueh as h~bùgc., sulfide, over non-polar gases, such as methane. In tests on natural gas st~arns, the l~ ,s 1~ ' ' ' L.~l-u~ sulfide/~ ' - sel~tiv;ly of 25-30 and carbon di~,- ide/ ".,l}.ane scl~livil~ of 7-8, which latter number wæ co.~;~.~ too low for ~1 r~l carbon dio~ide separation. The m~,.,ll"~lc was also shown to be physically unstable at feed pressure above about 1,290 kPa (170 psig), which, even if the carbon ~ ,lLd..c sel~livily were adequate, would render it ~ v;tablc for k .. ll;.. g raw natura1 gas streams. U.S. Patents 4,608,060, to S.
Kul~,..Jtl,i~ ~,a, and 4,606,740, to S. Kul~" ' . j~ and S.S. Kl~ rni, of Li's group at UOP, present ad~ ;o1~l data using the same ~pe of glycol-laden membranes as .lic~..cc~ in the DOE report. In this ~Ivo 95/11738 2 ~ ~ 4 ~ ~ 7 PCT/US94"2099 case, however, pure gas tests were performed and ideal hydrogen sulfid~,.~ c sclc~liviLes as high as 115-185 are quote~ It is ill~lillg to nde that these are 4-8 times higher than the later Ill~&J~ Ai mixed gas nu.,,b.,.~ quoted in the DOE report. The same cffect obtains for carbon dioxide, where the pure gas sclc~Livitics are in thc range 21-32 and the mixed gas data give sclc~livilics of 7-8.
S Similar in concept is U.S. Patent 4,737,166, to S.L Matson et al., which ~ OSf C an immobili7~d liquid ~ e typically contailung n-lllclllylpyllulidollc or another polar solvent in c~ k~se acetate or any other r r ~ polymer. Thc membranes and l~o~sscs ~ e~ in this patent are directed to selective l~bu~c~ sulfide removal, in other words leaving both the methane and the carbon dioxide behind in the residue stream. As in the UOP patents, very high hydrogen sulfidc/i ' - scl~;tivilk s, in the range 90-350, sre quoted. Only pure gas data are given, however, and the feed pl~;~aul~ is 793 kPa (100 psig). There is no .1;~ io.~ as to how the ..,c."l., ,~,s might behave when exposed to mlllsicc ,.-poll~"lt gas streams and/or high feed }"~ ,u,~s. Based on the UOP ~ , the mixed gas, high-pressure results might be ~I,c~t~i to be not so good.
U.S. Patent 4,781,733, to W.C. R~ et al., dc~"il~s results obth;-.~i with isn illt~
15 c~"~ih""~.,ll" - made by a pohJ ~ s.. reaction between a diacid-chloride- terminated silicone rubber and a diarnine. In pure gas e,.y~ at 793 kPa (100 psig), the membrane e ~ gel.
sulr~dc/i -' - scl~tivili~,s up to 47 and carbon d;u.ude/ ' - s~l~livilics up to ~0. No mixed gas or high-pressure data are given.
U.S. Patent 4,493,716, to RH. Swick, reports permeation results obt~ ed with a ~m~u,il~
20 ..,~ IC c~ g of a polysulfide poh~mer on a Gor~x (pûlytctl~-nuulu..~l,ylene) support. Only pure gas, low-pressure test cell p .~ bilily dah are given. Based on the reported p. .. ~.r ~b.~ c, which only give an upper limit for the methane y~ jlity~ the membrane appears to have a L~rd~ughl ~ulrldc/m~lhanc scl~livi~ of at lc~st 19-42 and a carbon ~ ' ~ selc~livity of at least 2-6. Some results show that the carbon dioxide p~, ....,al~ility i,.-"-,a3~1 after C.~IG .UI~i to }I,~dlugC n sulfide, which 25 might suggest an overall d~,~ asc in scl~livily if the membrane has bccome generally more p~
ough no ...t~ .r data that could confirm or refute this are cited.
U.S. Patent 4,963,16~, to I. Blume and I. Pinnau reports pure gas, low-pressure data for a Cû~ O~;t~ .IC CO~ g of a yol~ pol~-lh~ block copol.~ .- on a polyamide support.
Hydrogen ,~lfidc~ ane ~le~ iti~.i in the range 140-190, and carbon IiuAidc/~ llane scl~liviLcs in 30 the range 18-20, are quoted. Mixed gas data for a stream c~ -P oxygen, nitrogen, carbon dioxide and sulfilr dioxide are also quoted and ~ -~ in the text, but it is not clear how these data would compare with those for ~ Ihn~c- or l.~l.ogcn sulfide~ "i- .;..g mixed gas streams.

WO 95111738 21 7 4 ~ ~ 7 PCTtUS94tl2099 1 Despite the many and varied research and d~ ul,....1 efforts that this body of lil4. G
rG~ ,e.~l~,ce~ oseacetateIll.,..lb~ s,withtheirD~ adv ~ag~'~ andd;s..lv gçs remainthe only ".~,...b.~.c type whose ~.o~,L~s in k ..lI;.~g acid gas streams under real gas-field ol,e. ., c ~ are ~ s -''y well ~.dw~lùod~ and the only n.~,.llbl - type in co..-..,e.-,;al use for ,.."ovu,g 5 acid gas c~"".o,.cn~ from mç~hDnr U.S. Patent 4,589,896, to M. Chen et al., exemplifies the type of process that must be adopted to remove carbon dioxide and L~u,~ - sulfide from methane and other h.~l~u~, l~ns when working within the performD~n~e limitations of c~ osç acetate membranes. The process is directed at natural gas streams with a high acid gas content, or at streams from; ' x ~ oil .~cu ~ ~ .y (EOR) operations, and0 consists of a .n..~ g~ -,e..lb.~c s p~ution, followcd by fi ~ I;o~ n of the acid g as CUIIIIJUII~I~ and flashing to recover the hydrogen sulfide The acid-gas-depleted residue stream is also subj to further l~ n~.l to recover l.~/-l-u.~ . The raw gas to be treated typically contains as much as 80%
ormorecarbondio~ide,withl.~l.uge..sulfideatthe,.' ~,lylow,fewll.u --.kofppmlevel Despite the fact that the ratio of the carbon dioxide content to the L~ ug~,n sulfide content is high (about 400:1), 15 the raw gas stream must be passed through a of four ,..~..ll,-~.e stages, .~u.gc~ in a three-step, two-stage c~ u,~io.., to achieve good l.~d,ugc., sulfide removal. The goal is not to bring the raw gas stream to natural gas pipeline ~ ';u - but rather to recover relatively pure carbon dioxide, free from hydrogen sulfide, for fur~her use in EOR The target couc~ - of carbon dioxide in the treated 1.~d,. bùn~ stream is less than 10%, which would, of course, not meet natural gas pipeline :~lLnl~ d~.
20 The methane left in the residue stream after higher l.~d~ ll removal is simply used to strip carbon dio!~ide from hydrogen-sulfide-rich solvent in a later part of the s-r _ process; no methane passes to a natural gas 1 . ' - Despite the . ~ t~,~./ '';~tagç 111~ 1 " arra~gf Qf-QI in a .~ ,s~lative e, ahout 7% carbon dioxide is left in the L.~Lu~,~l~u rcsiduc stream after ylu~ g and about 12% h~.Lu.,~l,u,. loss into the permeate takes place.

2~ In su.. --a-y, it may be seen that there remains a need for hllylu.~d Ill~,lnblanCS and hl.y.
y-~s~s for 1 lling streams c~ .g meth~ne, acid gas c~ ; and water vapor.

SUMMARY OF THE INVENTION
The .,. . _ltio.~ provides i~ u . d membranes and h~ u . ~d l..~,..lIJ. ~.e y. u~ s for ~eating gas 30 streams ch ~ i..g l.~dlug~,.. sulfide, carbon dioxide, watervapor and t~ ~, particularly natural gas st~ams. The ~,.u~ rely on the availability of t~-vo membrane typcs: one, cPll ' ~: acetate, or a material wi~ similar ~Jluy~lics~ cllal~ ,.i~ by a mixed gas carbon d;uAidf~--~ ane scl~livily of about 20 and W O 95tll738 ~ 3 k 7 PC~rtUS94/12099 a mi!~ed gas hydrogen sulfidc~ h~ e sel~livily of about 25; the other an i,--~,-u-~ I--e---b-a-,c with a much higher mixed gas hydrogen sulfidc/~ ' sclc.,livily of at least about 30, 35 or 40 and a mixed gas carbon d;~,AidcJl..elhane selccLiyily of at least about 12. These ~.,l~liv;Lics must be ~ with gas st~ams C'~ ;Qg at least methane, carbon dioxide and l,~.,g~,., sulfide and at feed p.~ UI~,;. of at S least 3,551 Wa (500 psig), more ~ ~y 5,620 Wa (800 psig), most p.~,fc. bly 7,000 kPa (1,000 psig).
Ani,-"~l aspectofthei--~_lti~listhea., l~hilityof.. l.~ ~cwithmuchhigherhydrogen sulfidc/~ ' ~ scl~liYil;~,s than cellulose acetate. This ~r~.;dcs the fleYihility to choose between the Ill~,~llblallc wYith the higher carbon d;~JAidcJIl~Lllallc scl~liY;ly, in treating streams c~ -.;..g little hydrogen sulfidc relative to carbon dioxide; the --.,.,~t - with the higher hydrogen sulfidch.,.lhà"c 10 scl~liYilr, in treating streams conhining s--~ 1 amounts of l,~og~,., sulfide relative to carbon dioxide; and a mixed ."~."b, - configu- in treating streams in the . ~ cat~u"Y.
The a~ ' ' 'ity ofthe two -~w,l~ C types enables il~,àLlll~lt p-u c;,~,s b-l ~ d in terms of the two ~llcnlblancs, so as to op~ ; - any process attribute a~ ~I;.,gly, to be d~ A Based on the different~ ~Lics ofthetwo ~ -.r types, we have li~ l that it is possible, through 15 ~Ill~u~ c-l;--g to derme gas c~ ;1;o-- zones in which a palli-,~lalll-,dl.~ "A process is favored.
For ~A . 1 1" if it is the primary goal to methane loss in the ....,..,1" allc pe ~e it may be better to carry out the ll~,d~ using only the more l.~,Ow.-sulfide-selective Ill~ (allC~ only the more carbon-dioxide-sclc~ ,.llbl - or a mixture of both, d~,r .~ , on the parlicular feed gas u....l~s;1;~
Sirnilar dl tt,~ ..-;-- ~ ;O~c may be made if the amount of membrane area used is to be --:--;---;~ the costs 20 and energy of ~,4,.~.;~ ~ are to be kept below a target value, the l..~dr~g~.- sulfide collc~.-L ation in the is to be ...,. ~;--.; -1. the overall operating costs are to be reAuced, or any other membrane process attribute is to be the key design factor.
If a ~ - of the two Ill~..llbl ~lC types is to be used, the pr~f,. . ~,d co f~t;u ation is to pass the gas stream first through modules c~ h;-~:--E the one membrane type, then to pass the residue stream 25 from the first badi of modules through a second bank containing membranes of the otha l~pe. If the raw gas stream contains s;~-;fi~ of water, for e . '~ it is p,~fe._b!e to use the more l,~ O_.,-sulfidc-s_l~li~ _ .. 1.. --~. first. 'fhese h are not usually ~' ~g ' by water, and can handle gas streams having very high relative ~ ' up to _ Fu ~ o~e, the I are very ~e to water vapor, and so can be used to dehydrate the gas strearn before it passes to the second 30 bank of - ~ ~
Any l".,."I,--.~cs that can achieve the necessary carbon dioxidcJ~ h P and l,~I.oge., sullid~"~,ll~ sel~livili~s undermixed gas, high-pressure cQ-ul ~;OI~c plus provide c~"u,.~ ,;ally useful WO 95/11738 ,21~7 4 3 4 7 PCT/US94/12099 L~ ,...~. . .r."h~ ~lc fluxes, can be used. The most yl~ir~ ,d material for the more carbon-dioxide-scl~Li~, ... ~ .. ~l .. r. is cellulose acetate or its variants. The most p.. ,fc l~ material for the more Ly~llu~ sulfide-scle~,Li.~ nbl- --c is a polyamide-poly~,LI.~.. block copolymer having the general formula HO~ C--PA--C--O--PE--C ~}H
S 0 0 n where PA is a polyamide se~mt nt, PE is a polyether segment and n is a positive integer. Such yOI~/III~,.:.
are available c _ -'b as Pebaxl9 fi~n Atochem Inc., Glen Rock, New Jersey or as Vc~L~.Iidg) from Nuodex Inc., Piscataway, New Jersey.
In their most basic .-..~ f-;, the yl~sses of the ill~.-dion make use of a one-stage 10 ~ , design, if a single ".c..lb, - ~pe is ~ - d, and a two-step m~..,l" - design, in which the residue from the first step becomes the feed for the second step, if a cu",b;"~ltion of Illc.llblallc types is - ' ' ~ytiunally, two-stage (or more c~ .pl~ ~1) membrane conr.~u-dfiùns~ in which the p ~ o from the first stage bc~ c the feed for the second, may be used. This will both increase the concc.lfJ ~.t;on of h~.Lug~... sulfide in the second stage yl ~ and reduce the methane loss.
lS The ,.. c.. b. - process may also be cc~ -~ with one or more non-,.. ~,."l".u,c ~"uce~ ,es, to provide a ll~dhll~ll scheme that delivers pipeline quality m~th~ne, on the one hand, and that cu"~.~dtes and disposes of the acid-gas-laden waste stream, in an c~vln ~ f~ y a -r 1 1_ manner, on the other.
The y~ sl,S of the i..Y~,.dio.. exhibit a number of adv 1 compared with ~ ,Y;o~ly available acid gas ll~h--~,-d t~ ' ~!cgy. First, y-u. i~;ùn of d Ill~.~lltilallC with much higher scl~liv;ly for 20 hydrogen sulfide over methane makes it p: ' le, for the first time, to apply ",~,..IL ~ Lll~af~ lt e~.,;~ ly to gas streams cl.~ a~t~.~ ;~d by relatively high conc~,.d, - of L~u~ sulfide. Sccu~dly~
the y,u~ are much better at ~ ng gas streams of high relative humidity. Thirdly, it is 5~
possible to bring a natural gas stream into pipeline 5r " f - for all three of carbon dioxide, hydrogen sulfde and watervaporwith a single nembrane treatment. Fourthly, ~,._.~..~ .,.~g of the gas stream by 25 ,~,.,,ùv;,.g the carbon dioxide to a much greater extent than is actually nc~ssa y, simply to bring the hydrogen sulfide content down, can be avoided. Fifthly, much greater ~ yibility to adjust membrane ~x-~ g and perfo~nance p ~ is provided by the a~_ l ' 'ity of two types of ",.,..,~ ,s. Sixthly, the process can be o~ -A for any chosen process attribute by cA~ ;-.g the appropriate membrane mix to use.
BRIEF DESCRIPTION OF THE DRAWINGS
Figure I is a diagram showing zones in which palli~,ul&~ ",~,."I"~,~s should be used to separate h~ogen ~=~
~43~17 ~NO 95/11738 PCTIUS94/12099 ,, 11 sulfide and carbon dioxide from methane.
Figure 2 is a basic ~ ' - drawing of a one-stage ~ ,.nljl~ulC separation process.
Figure 3 is a graph showing the effect of water vapor on carbon dioxide flux through ce~ ose acetate 1ll~,.1~11l a-lCS.
5 Figure 4 is a graph showing the effects of h~nLogen sulfide and water vapor on the p- rO.... C4 of lose acetate Ill~ ICS.
Figure S is a basic 5~,h~ ;c drawing of a typical two-stage ..~ - ~ s~ udlioll process.
Figure 6 is a basic 5~,hr-.. ~;~ drawing of a two-step membrane separation process.
Figure 7 is a basic s~h -.. ~ drawing of a two-step/two-stage membrane s.,~.u process.
10 Figure 8 is a basic 5~,h. ~ ;r. drawing of a two-stage ~ l.ulC separation process with an auxiliary ..en.b.~.e unit forming a second-stage loop.
Figure 9 is a diagram showing zones in which particular ,..~,..,~r~.ncs should be used to separate L~ ughl sulfide and carbon dioxide from ' e, based on different Ly~Lu~_.l a.llrldc~...lllane scl~liv;li.,s.
Figure 10 is a diagram showing zones in which paficular membranes should be used to separate h~d~ug~,n 15 sulfide and carbon dioxide from methane, based on different carbon di~"~idc/~ ' - s-,l~livilies.
Figure 11 is a diagram showing zones in which particular membranes should be used to separate h~d u~
sulfide and carbon dioxide from m~th~ne, for different feed gas ~..~ ....~,;..

DETAILED DESCRIPTION OF THE INVENTION
The term intrinsic scl~liYily, w used herein, means the s.,lc~iviLy of the polymer material itself, ' ~ ' as the ratio of the F ~ tif'S of two gwes or vapors through a thick film of the material, as -.ca ..u~d with pure gas or vapor samplcs.
The term ideal scl~livity, as used herein, means the sclc~livil~ of a ~ 7, ~ _~ i as the ratio of the p~ ' litie5 of two gases or vapors through the IIIC~IIlJIallG, as IIICUS~ d with pure gas or 25 vapor r The terms mixcd gas s.,l~liYil~ and actual sel~livity, as uscd hcrein, mean the sel~liYil~ of a L~ " calculated as the ratio of the ~ - .~ ilitirc of two gases or vapors through the 1ll ,.llbl -IC, as ..ca..~,d with a gas mixture containing at least the two gases or vapors in ~ ctirm The i..~,.llion has several aspects. In one aspect, the il~ lliùn co.~ ,S for treating gas ~ixlul~,s c~ carbon dioxide in certain collc~ s, I~ llu~ sulfide in certain co.~.lll..liolls and .n.,~ ., to remove the carbon dioxide and h.~J~ug.,.~ sulfide. In another aspect, the invention CQ.~,-..C u~ g such membrane separation l,.ûccsi~,s in terms of a particular process wo 95111738 Pcr/uss4/l2099 2 17~3~7 12 attribute This ~l;..~;,;..g may be done to . ~ , the methane loss from the ".~,...~.~c process, to the Lydlu~ sulfide c~ , ,t.~ in the p .... ,t~. stream, or to provide the best fit bet~-veen the ,...,."~-ane process and a non-l"c."b,_.e process or ~IU~;.~_5 acting together as a "hybrid" process, for example In yet another aspect, the i~ tiùll cc ~ ,.llbl~l~.5 that ~ high hy~Lùg~
S sulfidch..cll-~c s.,l~;liv;lics when ~ ngçd with mixed gas streams under high pl.,;.~
The p~occsscs of the i..~_lltio.~ rely on the availability of two membrane types: one, u ll~llose acetate, or a material with sirnilar ll~u~Li.,s, chal~. t~,.i~ by a mixed gas carbon ~liu idc~n~,ll,~.c sclc~,liv;ly of about 20 and a mixed gas h~ù~.. ,ulGde/ll.elh~lc scle~,Liv;ly of about 25; the other a n~ ulc with a muchhighermixed gas Ly~Lug~ .JlGdc/lnclL~c scl~livily of at least about 30, 35 or 40 and a mixed gas carbon I;o.~id~,.eLh~,c sclc~,livily of at least about 12. These scl~livili~,5 must be ~L.CV~ wi~h g~s streams ~ ~ g at least ' e, carbon dioxide and hydrogen sulfide and at feed plCia~ of at least 3,551 kPa (500 psig), more p,cfi,.d~ly 5,621 kPa (800 psig), most l"~,f~,.~ly 7,000 kPa (1,000 psig).
The invention provides three forms of basic l-l.,ml,l - I-~id~ ,.-L process:
15 1. Using only the more hydrogen-sulfide-sel~li~_ ",~,lllb,_~c 2. Using only the more carbon~liu.~ide-scl&Li~ n,~,llll" -

3 . Using a c~ .h:~ ~ ;O-I of both types of membrane.
Based on the differcnt p,- ..~ lu~li~s of the two ,..~ types, we have d;s~ . .~,d that it is possible, through crr~ e to define gas co. .I~;l;o ~ zones most ~ to each one of 20 the~e threc types of basic plUCci~a~S. In perfonning the u~ vtc cc~ ... a specific process attributc is used as a basis for calculating thc ~..~1... ;~,~ of the gas c~ zones. It will be apparent to those of ordinary skill in the art that any one of many proccss attributes could serve as the basis for the c~lc~ inn Rt~ scnldlh~-~ non-limiting, ~ , It include melhane loss, ~ c arca, stage cut, energy c~n ~ l ;o~ ~ annual opaating costs, ~ c ~ -. residue c~ best match with 2~ other plU~i~aCS in the tlcdtnl~lll train, ~ ~i~ll~lll~ailiull ûf recyclc strcams, and sû ûn.
Loss of mcthanc is usually onc of the most i,u~,ull~ll factors in natural gas ~lu~f ~ g On the one hand, pipeline grade methane is the desLred product, and s~ losses of product have a s. ~l .cn.. .1; ~ adversc effect on thc process economics. On the other hand, large ~, iti--c of methane in the acid gas stream make further 1- ..ll;.,g and I~CG ._ly of any useful IJIudu~ i from this strearn much more 30 &fficult. As a general rule, a ~ r ~r..l natural gas L ~t~ ,.d process should keep ~ ' - losses during u f ~ .g to no more than about 10%~ and ~ ,f~,l bly no more than about 5%.
For ~;.. I.li~ ity, Ih~e~c~ most of the ~ and ~ s have been &rected to plU~:._5 wo 95/11738 ~ l 7 4 3 ~ 7 PCr/US94/12099 designcd to minimize methane losses, nl~hOv~h it should be ~pl. : ~ that the scope of the .~ Lion is intended to ~ cc any process design calculations done with the same goal, namely, defining zones applicable to the various l -u~;~U~g options made possible by the two lll~,,lll l ,c types.
We believe the concept of these zones, how to e~lrul ! them and how to use them, is new, and 5 will be useful in treating any gæ stream that cc,..",.;~cs mf?th~n?, carbon dioxide and l~ùgc~ sulfide.
Such streams arise from natural gas wells, from carbon dioxide miscible nooding for c- h~ d oil ICi~ U ~ ~,ly (EOR) and from landfills, for example. We believe that it will be particularly useful in the s..~t~,l....g of natural gas c4~ g acid gas c ....~
Referring now to Figure 1, this shows a typical zone diagram, with feed gas carbon dioxide 10 co~ L, dLio.~ on one axis and hydrogen sulfide conchlll dliùll on the other. The diagram was ~ ~ by running a scries of ~ llblallC separation c~u~ t~ simulations for h~,,ol}.~ti~dl three-c~ l,o~
(m~th~n~., carbon dioxide, L~nllugcn sulfide) gas streams of pul Li~,ula~ flow rates and c~ û~ c In all cases, the target was to bring the strcam to a pipeline ~i1~ - of 4 ppm h~/~hugh~ sulfide and 2%
carbon dioxide. The lll~.lllbla.l~. ~,.o~.lics were ~ccllmf-~ to be as follows:
MoreCO2-scle~Li~ ..-b-~ ,. Carbond;u,~idc/ ' -scl~li~ 2û
.. sulflde/ ' - selc~Liv;ty: 25 ~ flux: 7.5 x 10~ cm3(STP)/w..?-s ~,...Hg More H2S-sel~li~, ln~.n~,ane. Carbon !~-f ~ hdnc SClC~.livily: 13 2Q IIyd~u~ s.,lrld~ ' - scl&livil~; 50 Mcthane flux: 7.5 x 10~ cm3(STP)/~,...~-s c...Hg In each case, thc methane loss into Lhe pameate stream that would occur if a one-stage membrane s.,~ lion process were to be carried out was c~ t- A and was used to define zones of least methane 25 loss. As can be secn, Figure 1 is di-vided into four zones. In zone A, no treatment is required, because the gas already co~ c less than 2% carbon dioxide and less than 4 ppm Ly~Lu~n sulfide. In zone B, meth~ne loss is ,.. .;.u~ 1 if the more h~og.,,.-sulfidc-scl~li~, membrane alone is used. In zone C, methf~ne lOSS is ...:-.:-..; ~;1 if the more carbon-dio~cide-sclc.,li~, membrane alone is used. In zone D, methane loss is minimized by using a combination of the two membrane types. The zones are calculated 30 based on the l~ ll'til ulC scl~livity and their exact position will change if the ~ f~ selc-,livity rh~nges Figures 9 and 10 show the change in the BID boundary for Ly~llugc.l sulfidc/lll~,lhdnc scl~livili.,s of 30, 40 and S0 and for carbon d;o.~it~..~ ane scl~tiv;li~s of 10, 13 and lS.

wo 95111738 Pcr/uss4/l2099 %~1.7~L3~ 14 The zone diagram may be used directly to d~,t~- r the best type of Ill, llblallc to use for a specific separation by reading off the zone into which the feed C<~ ~ 7ilioil fits.
Another way to use the diagram is to define COIl ,c~ ' on bands that can serYe as C~ linP~ in selecting a ~"~LI process. Again referring to Figures 1, 9 and 10, we have disw . w~ that, as a guide, S three carbon dioxide con~.lLIatioll bands may be defined, thus:
1. (a) If the feed gas to the ~ C system contains less than about 3% carbon dioxide to less than about 10% carbon dioxide and more than about 10 ppm h~og_., sulfide to more than about 300 ppm II~Ug~l sulfide, with the lower cnd of the carbon dioxide range Cull~ g to the lower end of the hydrogen sulfide range (<3% carbon dioxide; > 10 ppm l~JI uc,~,., sulfide) and the upper end of the 10 carbondio?~iderange~,~ dingtotheuppercndofthel,y~c,~llsulfiderange(<10%carbondioxide;
>300 ppm hydrogen sulfide), then thc most favorable process, in terms of ...;-.: ..;~;- g methane loss, is carried out using the more hydrogen-sulfidc-scl~li~ lC only.
(b) If the feed gas contains less than about 10% carbon dioxide to less than about 20% carbon dioxide and more than about 300 ppm hydrogen sulfide to more than about 600 ppm hydrogen sulfide, 15 with thc lower end of the carbon dioxide range co,.~ g to the lower end of thc L~d~u~ l sulfide range (<10/O carbon dioxide; >300 ppm l.~,d~o~n sulfide) and the upper cnd of the carbon dioxide range C~ r ' g to theupperend ofthel,~ugcn sulfide range (<20% carbon dioxide; >600 ppm hydrogen sulfide), then the most ~ le process, in terms of ... .~ ;--g methane loss, is carried out using the more Lydlu~n-sulfide-selcc1i~ bl - only.
(c) If the feed gas contains less than about 20% carbon dioxide to less than about 40% carbon dioxide and mo~e than about 600 ppm l,~l~u~ , sulfide to rnore than about 1% hydlug~,,l sulfide, with the lower end of the carbon dioxide range C411~_r ~- g to the lower end of the L~dlU~ I sulfide range (<20%
carbon dioxidc; >600 ppm h.~oc,_l sulfidc) and the upper end of the carbon dioxide range c~ dillg to thc upper end of the h~ sulfide range (~40% carbon dioxide; >1% L~U~11 sulfide), then the most f~vol ~le process, in terms of ~ --g methane loss, is carried out using the more L,~/dlUg~.ll-sulfide-scluli~ l.u.C only.
Also, three h~d~UgC/n sulfide concentration bands may be defined, thus:
2. (a) If the feed gas contains less than about 5 ppm L~dIU~II sulfide to less than about 50 ppm L~llUg~,.l sulfide and more than about 3% carbon dioxide to more ~an about 15% carbon dio~ude, with the lower end of the carbon dioxide range C~ on~g to thc lower end of the L~U~n sulfide range (<5 ppm hydrogen sulfide; >3% carbon dio~cide) and the upper end of the csrbon dioxide range ~ .li.,g to the upper cnd of the h~JJogcn sulfide range (<50 ppm hydrogen sulfide; >15% carbon wo 95/11738 2 ~ ~ ~ 3 ~ 7 PCT/US94/12099 ~s dioxide), then the most f..~ le process, in terms of ~;-.;...;,;-.g methane loss, is carried out using the more carbon-dioxide-s~ C only.
(b) If the feed gas contains less than about 50 ppm l,~Lo~,., sulfide to less than about 250 ppm hydrogen sulfide and more than about 15% carbon dioxide to more than about 50% carbon dioxide, with 5 the lower end of the carbon dioxide range cUl~ U~ t;~ to the lower end of the h~d-ugh~ sulfide range (<50 ppm L,~o~.l sulfide; >15% carbon dioxide) and the upper end of the carbon dioxide range C(lll~_r ~- g to the upper end of the l,~dlug.,.. sulfide range (~250 ppm L~IIUK~I sulfide; >50% carbon dioxide), then the most Çh~ process, in terms of minimizing methane loss, is carried out using the more carbon-dioxidc-sele~ c membrane only.
(c) If the feed gas contains less than about 250 ppm hydrogen sulfide to less than about 500 ppm hydrogen sulfide and more than about 50% carbon dioxide to more than about 85% carbon dioxide, with the lower end of the carbon dioxide range CUII~ to the lower end of the II.Y~UgC~I sulfide range (<250 ppm hydrogen sulfide; >50% carbon dioxide) and the upper end of the carbon dioxide range cUII~ to the upper end of the l.~ u~,.. sulfide range (<500 ppm hydrogen sulfide; >85% carbon lS dioxide), then the most fa~ul ' le process, in terms of ~ methane loss, is carried out using the more carbon-dioxide-scl~li~ only.
Also:
3. For feed gas c~ n~ ~;O~c outside the ranges sl,c~ ~ in points 1 and 2 above, the most fa.~,l ' !e process, in terms of ,..:-. -..;,; ~g methane loss, is carried out using a co~n~ of the more LyLug~,.--20 sulfidc-scl~li~ and the more carbon-dioxide-scl~li~, manbranes.

For in~t~nCÇ, using eitha the zone diagram itself or these conc~.lhd~ioll bands:(i) If the carbon dioxide content of thc stream is 4.5%, and the L~ .Lu~h~ sulfide content is 1,500 ppm, the more hydrogen-aulfidc-sel~li~ -~.c only should be used.
25 (ii) If the carbon dioxide content of the stream is 4.5% and the 1.~ .. sulf~de content is 7 ppm, the more carbon-dioxide scl~li~, membrane only should be used.
(iii) If the carbon dioxide content of the stream is 4.5% and the L~u~ l sulfide content is 25 ppm, a c~ .llbl - system should be used.
(iv) If the carbon dioxide content of the st~n is 7% and the l.~u~. sulfide content is 10,000 ppm, the 30 more h~u~cn-sulrldc-acl~li~, mambrane only should be used.
(v) If the carbon dioxide content of the stream is 7% and the Ly~hu~n sulfide content is 2 ppm (already within spcc.), the more carbon-d;o~i~-scl~li~_ membrane only should be used.

wo sstll738 ~ 1 7 ~ ~ ~ 7 Pcr/Uss4/12099 (vi) If the carbon dioxide content of the stream is 7% and the hydrogen sulfide content is 50 ppm, a cu.nb;..dlion l..~,.-.I,-~c system should be used.
(vii) If the carbon dioxide content of the stream is 10% and the l,.~l~og~". sulfide content is 1,000 ppm, the more hydrogen-sulfide-scle~,Li~ , only should be used.
5 (viii) If the u~rbon dioxide content of the stream is 10% and the L~.LuO~n sulfide content is 20 ppm, the more carbon-dioxide-selc.,li~ ,...b- - only should be used.
(ix) If the carbon dioxide content of the stream is 10%, and thc h~&uo_~ sulfide content is 100 ppm, a co.~ ,m~, system should be used.
(x) If the carbon dioxide content of the stream is 16% and the L~Luo_~l sulfide content is 7,000 ppm, the 10 more hydrogen-sulfide-selective l-,e..-b-_ ~ only should be used.
(xi) If the carbon dioxide content of the stream is 16% and the l.~c,gc.. sulfide content is 8 ppm, the more carbon-d;ohidc-selective l..~,..L. 7 only should be used.
(xii) If the carbon dioxide content of the stream is 16%, and the h.~l~uo~ sulfide content is 250 ppm, a co...bination I.~...L. .~, system should be used.
15 (xiii) If the ca~bon dioxide content of the stream is 25% and the h, ~L UO~,.. sulfide content is 10%, the more hydrogen-sulfide-sel~li~c n..,.nbr ac only should be used.
(xiv) If the carbon dioxide content of the stream is 25% and the h~d~og~,.. sulfide content is 50 ppm, the more carbon-d;u~idc-s~ ~ only should be used.
(xv) If the carbon dioxide content of the stream is 25%, and the 1.~UO~I sulfide content is 500 ppm, a 20 cc...L...alio.. ~ L-al~, system should be used.
(xvi) If the carbon dioxide content of the strearn is about 50-60% or more, the more L~Lug~ sulflde-scl~li~_ ,...,..lbr - should not be used alone, no matter how high the l..~ùg~., sulfide content.
(xvii) If the carbon dioxide content of the stream is 40% and the l.~d,u~;~,.. sulfide content is 120 ppm, the more carbon-dioxide scl~ti~, membrane only should be used.
(xviii) If the carbon dio~ide contentofthe stream is 40%, and the i~ u~ sulfide content is 2,000 ppm, a u~ h; ~ ~ jon ,.. ~.. I,.. e system should be used.
(xix) If the h~uo~ - sulfide content of the stream is about 600 ppm or more, the more carbon-dioxide-s,~ membrane should not be used alone, no matter how high the carbon dioxide content.
(xx) If the carbon dioxide content of the stream is 70%, a combination membrane system should always be uscd if the h~d~uO~,n sulfide content is above about 500 ppm.
The ~ :o~l of the zone diagram and the specific ;-.~r~ - rS of what it teaches for twenty different gas co..~l~o~i~;o-~c is dclil" ~t,ly fairly lengthy, so as to cover e pl~ in the mid-ranges and ~ 74~7 WO 95/11738 Pcr/uss4/12099 near the cdges of the bands and zones.
Another way to express the 1~ ~l.;..g~ of the u~ tion is simply to define single limits for the carbon dioxide and ~ -UgCll sulfide concc.,l atiOnS that are best treated by different types of .n.,...b --G.
This approach gives a less accurate result in any ~lLv;dual c..~,w~Ldnce than the zone or band S app-uacllc~, but gives a broad guide that is useful ul~Li~, of the particular process attribute that is of most concern. Spifir~lly:
1. If the carbon dioxide content of the stream is less than about 40% and the 1., ~ og~,. sulfide eontent is more than about 6,000 ppm (1%), the more hyd~uOI,-. s~llrldc-scle~ lallC should be used.
2. If the carbon dioxide eontent of the stream is less than about 20% and the l.~.Luo~.l sulfide eontent is 10 more than about 500 ppm, the more hydrogen-sulfide-sel~lh,_ --c.--l,-,u.c should be used.
3. If the carbon dioxide eontent of the stream is less than about 10% and the LyJIogw~ sulfide content is more than about 10 ppm, the more l.~.huoc.,-s.llrldc-sclc~Li~_ membrane should be used.

4. If the L~/JIùg~l sulfide eontent of the stream is less than about 25 ppm and the earbon dioxide eontent is more than about 10%, the more earbon~l;o~idc-scl~li~ only should be used.
15 5. If the hydrogcn sulfide content ofthe stream is loes than about 100 ppm and the earbon dioxide content is more than about 15%, the more earbon~l:oAidc-sel~li-_ membrane only should be used.
6. If the earbon dioxide eontent of the stream is in the range about 5-20% earbon dioxide and the hydrogen sulfide content is in the range 10-1,000 ppm, a eombination ....,..ll, 7 system may be used.
7. If the carbon dioxide eontent of the stream is in the range about 10-25% earbon dioxide and the I.ydloO~.. sulfide eontent is in the range 50-5,000 ppm, a eombination ll.~,.--I,-a.-c system may be used.
8. If the carbon dioxide content of the stream is greatcr than about 25% earbon dioxide and the l~ gc~
sulfide content is greater than about 200 ppm, a eombination membrane system may be used.
9. If the carbon dioxide eontent ofthe stream is greater than about 40% earbon dioxide and the h~.huo_.-sulfide content is greater than about 600 ppm, a combination membrane system may be used.
If a c~ of the two membrane types is to be used, the pl~ config-u~lio - is to pass the gas stream first through modules c~ ag the one membrane type, then to pass the residue stream frofn the first bank of modules through a seeond bank c ~i~ .,,.,,.ll~.~u.~,s of the other type. The order in whieh the mcmbrane types are e'~''4 -~ d by the gas stream ean be ehosen according to the ~ - s of the applicatia~L If the raw gas stream eontains signifieant amounts of water and l.~.,gm sulfide, for 30 1 , 'e, it is p~ ' 'e tousethemorel.~huO_~ s~rdc-3clc.,li~_ lll.,mllanc first, sinee eel' ' s aeetate ...braucs have been shown to lose both scl~liv;ly and ~ s b. -'ly if exposed to combinations of water vapor and h.~ JIuOcn sulfide. They also do not withstand relative 1-- -- - ~ Q above ;
wo 95111738 2 ~ 7 ~ 3 ~ 7 18 Pcr/uss4ll2o99 about 30/O very well The poly ' pol~llh,. block ~pol~...~,. ..~ ..11., ~... ~ that are p.~,f.,..~d as the more l,~ugcn-sulfidc-~selecli~_ membrane, on the other hand, are not usually ~ ~ by water or l~ ug~,.-sul~de, and can handle gas streams having high relative h . . .:.l;l ;. c such as above 30% RH, above 90%
RH and even s.~tul~Lion. Fu~lh~ llu~G~ the lll~ l~l.,S are very p ..,~1 !e to water vapor, and so can be S used to dehydrate the gas st~eam before it passes to the second banlc of ~ ' "! If Lu~liLty and l.~.l.u~
sulfide content are not issues, and no othGr factors that affect only one of the m.,..,l,. types are at work, then the total methane loss into the F streams and the total .nc...l,.~.c area rGquirGd to perform the s~pr .~tioll should be ecc-nti~lly; L 1~ ~P ~ of the order in which the --~ b- IcS are positinnrA
Any.. l.. a cthatcanachievethenccc;.. - ycarbondioxidelmethaneselc~liy;lyandl~yd~u~
10 sulflde~ Lllullc :,~,l~liviLy, plus cc l....e..,;ally useful t, ~ . h~.c fluxes, can bc used.
Preferably the Ill~ lal~S should be chaua~,t~,.i~ by ~o~ bla~c methane fluxes of at least 1 x 10~ cm3(STP)/c..l7-s.,MHg, most p-~f~.~bly by h ~ b~allc methane fluxes of ât least 1 x 10~5 cm3(STP)I~ s ~,...Hg.
For the more carbon-d;oAidc-scl~~ , the ~.~f~ d ...~,...1,. are the ce~ ose 15 acetate .~ ~ that are already in use. Other c ' ' s include different c~ cr d~..;v ~., " such as ethylrelllllnc~ ylcclllllocP r~u~lhllose and palli~.uLIly other cP-llnlose esters. Oih~ .;sc, ,--~,.-.l,- .u-~,s might bc made from polysulfone~ h~ l rnn~ polyamides~ poly ~ , pol~ _lh~,. ".ll~, polyacrylonitrile, polyvinylalcohol, other glassy n.at~..;als or any other ~, u~,.;..te mntenql Usually, glassy md~.;als have enough ~ ! strength to be formPd as integral ~.. ~.t~ic membranes, the 20 p. u l~L~I of which is well lmo -vn in the art. Thc i..~ _.ILu.. is not intended to be limited to any particular -a..e material or ,..~,n~.~.c typc, howevcr, and . pr any ..~....I,.anc, of any mntPn~l, that isc~pableofmeetingthetarget~ ,.u~.li~ p forexarnple,h.,...n~--l~ membranes, cu...~ ile ,..~...b. - s, and ~ ,.llb- incorporating sorbcnts, carriers or p~ . -- a.
For the more L~ u~ sulfide selc~ _ membrane, the most ~.cf~ d membranes have 25 hyd,ophilic, polar elnctnme~ic sel~liv~, layers. The mobility selc.liv;l~ of such matcrials, ~lthnll~h it favors hydrogen sulGde and carbon dioxide over ' e, is modest ca--.r ~d to glassy ;als.
Because the ll~lllllall~, is l.~ l-ilic and polar, howcver, the sorption sel~livity strongly favors Ly~hùæcn sulfide, carbon dio~ide and water vapor over non-polar h~dl~rh~ ~ gases such as h,~u~,.., ' ~, propane, butane, etc. Although the sclc~l;vily of such mak:rials is affected by swelling in ~e p-~e.~cc of 30 c ' - ~ y~Y~ t`;, wehaveJ;s~.~ ithatL~nllug~ lfid~..~.ll-~-escl~Lv;li~,sofatleast30or 35, 5C...~,I;.... c at least 40 and so.--- 1;----~ 50, 60 or above can be maintained, even with gas n..~ ,s con~ .g high acid gas concentrations, even at high relative h~-l.li~y, and even at very high feed ~743~7 yl~ul~ up to 3,551 kPa (500 psig), 5,621 LPa (~00 psig), 7,000 kPa (1,000 psig) or above. These are unusualandveryusefulylulfwLi~. Thesey.~fwli~renderthe,l-~ . sunusuallysuitablefortreating natural gas, which often contains multiple c ~ ~ ; has high humidity and is at high pressure.
Preferred membrane materials exhibit water sorption greater than 5%, more preferably greater than 10%,

5 when exposed to liquid water at room t~.lly~,.dtul~. Particularly y-efc.,~ are ~ t~,d or block copolymers that form two-domain SIIU~1UI~S~ one domain being a soft, mbbery, L~d~oyl~,lic region, the other being harder and glassy or more glassy. Without wishing to be bound by any yalli~,ul~ theoly of g~c ha c~ we believe that the soft, rubbery domains provide a ylCf~,~e~l1idl Yd1hV~a~ for the L~r~L~,~., sulfide and carbon dioxide C~ pO~ k~ the harder domains provide ~--rf~l~n~-~r~l strength and prevent 1 0 CA~ . swelling, and hence loss of scle~,livil~, of the soft ~lomginc Polyether blocks are yl~f ~l~d for forming the soft flexible d~ most yl~,L~l~ly these blocks i"~-yO~ polyethylene glycol, polyte~,a."~,ll.ylene glycol or polypropylene glycol, to increase the sorption of polar ~lecvles by the n,~ "b,~,e material.
One specific example of the most p,~,f, ,~ nl~ alle materials that could be used for the more 1S hydrogen-sul~deselective..~ .1... r ispoh~ ' blockco~ havingthcgeneralformula HO~C--PA--C--O--PE--C ~H
O O n where PA is a polyamide se~m~ ~t PE is 8 pol.~ - segment and n is a positive integer. Such pol~
20 are av~ ~,u"w~,;ally as Pebax~) from Atochem Inc., Glen Rock, New Jersey or as Ve;,t~...lg) from Nuodex Inc., Pi~dt..~ _" New Jersey. Thc polyamide block gives strengLh and is believed to prevent the ",e."l"a.,e swelling eAlX;~ _Iy in the y~ .lcc of water vapor and/or carbon dioxide.
OLher spccific examples include pol~_ll.cr- and pol~_st~,.-based ~I~ u~ ' Representative polymer fu~ and recipes are given, for example, in U.S. Patent 5,096,~92, in which the 2~ copolymers are made by first ~ iu,g a yl~ ( by co...k. - .g simple diols and aliphatic or aromatic ' I,u~lic acids with an excess of diacid to prepare diacid ~ blocks, then chain e ~ ..l;..g these with ~"ul"iatcly selected pol~,u~ .,c or pol~ .,e glycol sc~-.~ C
Usually, rubbery materials do not have enough mechanical strength to be formed as integral &~ r..lnlf.,llic~ .lllbla~,S~ but are instead i~w~ d into ~ ;t- membranes, in which the rubbe~y 30 scle~,livc layer is s~po ~ on a upo,ous substrate, often made from a glassy polymer. The dtiù~l of co...l~o~ ",e..l~, is also well known in the art. It is ~ ---o~ly thought that rubbery ~---r ' Illwllllla les do not withstand high-pressure op-, ation well, and to date, such membranes have wo 95/11738 ~ ~. 7 4 3 ,~L 7 20 PCT/US94/12099 not been generally used in natural gas ~ .h~ , where feed gas l~aaUl~.a arc often as high as 3,551 kPa (500 psig) or 7,000 kPa (1,000 psig). We have found, however, that C~uln~ait.e m~,ln~ CS~
with thin enough rubbcry scle~,ti~, layers to providc a tl ~ c methane flux of at least 1 x 10~ cm3(STP)/~.In'-s ~,mHg, can be used s ,t;~r~ 4~. ;ly at high feed pl~,aaUl~a and not only ..-J~ *
S their integrity but continue to exhibity high selc~liv;Ly for L~ugf n sulfide over .... Ih .~
In their most basic .,---ho~ the plU~aSCS of the invention make use of a one-stage II~WIII1I design if a single n.~,.,l,. - type is ' i and a two-step membrane design, in which the residue from the first step b~cu..l~,s the feed for the second step, if a cv~ L~1 ;on of membrane types is -' ' It will be apparent to those of ordinary skill in the art that more soph---- i P...h~l~ f ."~i arc 10 possible. For ,~ . ', a two-stage (ormore complicated) u~lll~l - configuration, in which the p~,lllcdte from the rlrst stage becomes the feed for the second, may be used to further enrich thc acid gas content of the pe,lll~,dle stream and to reduce methane losses. It is Glv;_..B~ that a two-stage ~ l IC
configl-ration, using like or unlike Ine..l~ types in the two stages will often be used. In such ~ i, the residue stream from the second stage may be l~.uLAt~ for further tl~,aLInf nl in the 15 first stage, or may be passcd to the gas pipeline, for ~? , t7 In one-stage cOnrlg~ll ~n~A, the residue stream may also be a~ll);e -~ l to further l..c,lll~l~.e wll. Bothp~ tr andresiduestreamsmaybes ~jGC ~to r~ non-lll~ at~
such as in an amine plant, to bring it the residue stream to pipeline ~r - '' ~ - 7 for, ~ Given the diversity of flow rates, c,- .. ~ n~ and locations of natural gas wells, it is ~ v;s;ollcd that the membrane 20 sep,l.dli~l process will often form part of a hybrid tl~idtnl~lt scheme that delivers pipeline quality r ' e, on the one hand, and that conc~ ,ates and disposes of thc acid-gas-laden waste stream, in an ellV;IU~ 9-lly ~ J~ manner, on thc other.
In the zone cgl- -'- the target pipelinc speeir~ ' for the treated gas was s~cllmpxl to be no more than about 2 vol% carbon dioxide and 4 ppm h.~og~". sulfide, which is typical pipeline 2S ~ ;.... Ilo ..~ p~ ..g on the destination of the gas and specific ~la to which the gas is subject, it is believed that a carbon dioxide content below about 3 vol% and a h.~ "l sulfide content below about 20 ppm will bc a ~ in many c;~ ;""~
The plU~âi~,5 of the Ul~.lltiOII exhibit a number of advantages cu,ll~ ,d with pl~Luualy available acid gas l.~ 4 t~ ~. First, p~V;a;O~l of a ~ ln le with much higher sele~Lvity for 30 ~ IIIU~II sulfide over methane makes it pos ' k, for the first time, to apply membrane L,.,d~llle,ll e~ "dlyto gas stleams cl~t~;~d by 1.1 ~ high ~,un~.lll..t;ùlla of L~LUg.~ll sulfide c~ r ~d to carbon &oxide. This expands the range of utility of lu~,llll,l - s ,~ 'ly. Since ~ ,nl~

wo 95/11738 ~ ~ ~ 4 3 ~ 7 PCT/US94112099 systems are light, simple and low~ '`4 Cullly ~,d with amine plants, the e-~h ~ d ability to use ..~,..lI,. _-c.s as a l-~dtl....-l option r~ il;t ~ ,s the ~ - l ,lo;l ~l ;nn of gas fields off-shore or in remote lor ~ ..c Secondly, the yro~csa~s are much better at ~ ~ing gas streams of high relative Lu If.lily, so that less yl~ ,dhl~ ll of the raw gas strearn is n occ~ .~ y. Thirdly, it is s~....- ~;- -- ~ possible to bring a natural gas S stream into pipeline sr ~ c~ for all three of carbon dioxide, Ly~ug.,.l sulfide and water vapor with a single ll.w,L-~ ,dh.~ll. This is a very i~llyl-JlL.II feature, which makes the pr~sa~.~ of the il~ ILio.l clearly more attractive than using one process for d~lh~d~ iull~ a second for carbon dioxide removal and a third forhydrogen sulfide removal. Fourthly, u. ~.yluc~ ~;--g of the gas stream by .~.nuvi~g the carbon dioYide to a much greater extent than is actually ne: ~, simply to bring the L~IIU~ I sulfide content 10 down, can be avoided. Fifthly, much greater neYihili1y to adjust ..~,..-l,-~e oy~,~dlillg and perfc - -pal~ ,lc~a is p,uvidcd by the availability of two types of ...~,..L-_..,s. Sixthly, the process can be ~y~ for any choserl process attribute by c~ the ~pluylid~ c mix to use.The i"~.,tioll is now further illustrated by the following ç~ ,lcc which are intended to be illua~ ~ti ~ _ of the i.- . ~-tiu.,, but are not intended to limit the scope or underlying I r I of the i..~ ~,nlio 15 in any way.

EXAMPLES
The e.~ lcs are in seven sets.
SET I
20 F.Y:mnplec 1-10 are compara~ive examples that illustrate the perfom~~nr~ of various glassy and rubbery polymers exposed to acid gæes under a variety of co .~ ...c E~cample 1. Pure gas l..e.,~ ,."~ . Poly;.";de "~,."I.,~,cs of two grades (a) A three-layer ~IIIpGait~. membrane was prepared, using a mi~ U~IUUa polyvinylidene fluoride (PVDF) support layer. The support was first coated with a thin, high-flux, sealing layer, then with a 25 sele~ , layer of polyirnide (l~r ~ Grade 5218, Ciba-Geigy, Ha~ll.ull,e, NY). Me..~ stamps were mounted in a test cell and thc y~ yrùy~lics of the membrane were tested with pure carbon dioxide and ~-vith pure .~ h '~G at a feed pressure of 448 kPa (50 psig). The results are listed in Table 1.
(b) A three-layer ~ membrane was yl~ya~,d, using a ll~ieluyoroua ~l~vi l~lid("u fluoride (PVDF) support layer. The support was f~rst coated with a thin, high-flux, sealing layer, then with a 30 selective layer of yol~i""de (custom-made 6FDA-IPDA). Membrane stamps were .- --,- -- -t-~l in a test cell and the pc---~ ~;o~- ylope~tics of the membrane were tested with pure carbon dioxide and with pure mP~h"-lP at a feed yl ~ ;~a.~lc of 448 kPa (50 psig). The results are listed in Table 1.

WO 95/11738 Pcrluss4/l2099 3 ~ 7 22 E.~ ,lc 2 Mixed gns Illcaaw~,.ll~,.lls. Pol~;...;dc --~,--L-~-~,s of two grades(a) Three-layer c- ",~po~ le membranes as in r--- p~ 1 (a) were tested with a gas mixture c ~ g of 800 ppm hydrogen sulfide, 4 vol% carbon dioxide, the balance ' -. The feed pressure was2,793 kPa ( 390 psig). The results are listed in Table 1.
S (b) Three-layer cu.. I~ ,.,.1,. as in ~ 'e l(b) were tested with a gas mixture c~ c;~l;. g of 800 ppm l-~ u~p, - sulfide, 4 vol% carbon dioxide, thc balance meth~lt~ Two feed p. ~ aw .,i., 2,807 kPa (392 psig) and 4,890 kPa (694 psig), were uscd. The results are listed in Table 1.
F , !e 3 Pure gdS ~..ca~w~ a PTMSP 1ll~
A cu...pûs;lc ...~,..11,. - was ~ -r ~ ~d by coating a polytrimethyl-sil~lp~ --c (PTMSP) layer 10 onto a polyvinylidene fluoride (PVD~) support .,~...l" ~. Mcml,l anc stamps were . ~ ~., ,ted in a test cell and the pe....e,llion },.up.,.li.,s of the ,..~,.,L, - were tested with pure carbon dioxide and with pure methane at a feed pressure of 448 kPa (50 psig). The results are listed in Table 1.
Example4 Mi~cedgas..eæw~,---~,.lls PTMSP,..."..I- ae COIllpOaitC ~ ,.llbl.,~,s as in r . ~- 3 were tested with a gas mixture co~c:~ g of 800 ppm L~d~ùg~ sulfide, 4 vol% carbon dioxide, the balance m~th ne The feed pressure was 2,793 kPa (390 psig). The results are listed in Table 1.
Example 5 Pure ~eas mcasul~ ..lla Silicone rubber "-~,.-,b-a--e A c-....l.os;le ~ b~ e was pl~,~Jal~l by coating a silicone rubber layer onto a llf.clu~luua support ,n.,mb. - Mc,llblalu stamps were molmtKI in a test cell and the pc.lll~,dliull ~,-ul.c.lics of the 20 membrane were tested with pure carbon dioxide and with pure methane at a feed pressure of 448 kPa (50 psig). The results are listed in Table 1.
Example6 MixedgasIllcaa~ lb. Siliconerubber--~,---b-Co...l.~ l1)l -- as in Example 5 were tested with a gas mixture co~ l; g of 650 ppm h~ugw~ sulfide, 4 vol% carbon dioxide, the balance ' - The feed pressure was 759 kPa (95 psig).
25 The results are listed in Table 1.
E.~ml,lc 7 Pure gas ~ ,,. ,w~..-.,.ltS Pol~ulad;~,,.e ...~....b-a.,e A c~ e ",~,.nl,. - wæ prepared by coating a pOlyln e (S~ ';e Polymer Products, Ontario, NY) l~yer onto a PVDF support ..~ Me.,ll" a..e stamps were .~ ~o. -t ~ in a test cell and the ~ ,o~.li.,s of the ~ - wae tested with pure carbon dioxide and with pure methane 30 at a feed pressure of 50 psig. The results are listed in Table 1.
r , ln 8. Mixed gas ,ne&~ . Pol~ ' - o ~ .a..e C~ o~ ; manbranes æ in r ~ lo 7 wae tested with a gas mixture c~ g of 800 ppm ~7~34'7 ~o 95/11738 Pcrluss4/l2099 I-~.l,ug~n sulfide, 4 vol% carbon dioxide, the balance methane. The feed pressure was 2,821 kPa (394 psig). The results are listed in Tablc 1.

Pc,~ al;on l~u~ L~,s of Various Glassy and Rubbery Polymer Mc..~.

L ..... - ~; X~

10I(a)(pwegas) 448 (50psig) - 23.0 1.23 - 18.7 2(a)(mixedgas)2,793 (390psig) 16.5 18.3 1.73 9.5 10.6 l(b)(puregas) 448 (50psig) - 156 2.47 63.2 2(b)(mixedgas)2,807 (392psig) 25.1 51.5 2.4010.5 21.4 2(b)(mixedgas)4,870 (694psig) 24.8 47.9 2.51 9.9 19.1 153(pwegas) 448 (50psig) - S24 281 - 1.9 4(mixedgas) 2,793 (390psig) 101 72.4 30.6 3.3 2.4 5(pwegas) 448 (50psig) - 41.4 10.7 - 3.9

6(mixed gas) 759 (95psig) 107 50.8 15.5 6.9 3.3

7(puregas) 448 (50psig) - 119 21.2 - 5.4 208 (mixed gas)2,821 (394 psig) 298 110 35.6 8.4 3.1 The highest ~I~L~,ity forhydrogen sullide ~vermelhane was only 10.5, which was ~ ~ with a polyimide ....,n.br - at about 2,862 kPa (400 psig) feed presswe.
Example 9. 13ch.lvior of Ce~ ose Acetatè Mc..ll,.~cs in the Pl~,sc..~ of Water Vapor This CO-~ ~a~i~C e rl- is from the report by N.N. Li et al. to the Dc~ of Energy ("Mc...... ll,-~-c S~.U.IL;UII P~uc~s~. in lhe pt tl. ' ' Industry," Phase II Final Report, Sept~ .. I r,-1987).
Li et al. ~ ~ the effect of water vapor in a feed gas stream of carbon dioxide on I ' - flux.
Figure 3, taken from the report su.. ;~i. their data. For relativc 1 l ~ of 10% or less, there is no ~I,.~iable effect on the carbon dioxide nux. For relativc 1 ' in the range 18-23%, the flux 30 dc~.~d 30% ~ ~ to the dry gas flux, but ~ d when the feed was s~.it~llcd back to dry gas.
Forrelative l .. "1;.;~ c of30%andhigher,thefluxdeclinewasfoundtobelarge,rapidandi~ ;ble.
E~ , 'r 10. I3ellavior of Cellnlose Acetate Membranes in the Pl~,s~ of Il~ ~o~,.. Sulfide and Water .

8 Pcrluss4ll2o99 ~ 2~3~7 24 Vapor This example is also taken from the Li et al. report. Figure 4 S.lllll-l~;~S the data. Hydrogen sulide has a negligible cffcct on ..,~"nl~.- -e perfonn~nre if the fecd gas is dry. If both L~JIugC.I sulfide and watcr vapor arc present, howevcr, the l- ~ c flux is s_bs -'ly reduccd. Li et al. cn~
S that the ~,r~c ,sing of streams c~ -.g both high COllC~ t~_ ~ of L~oghl sulfide and water vapor must be avoided with c~ ose acctate n.."..~ es.

r , ~ and 12 show the perfnnn~nr~ of pol~ ' pol~ , membranes exposed to pure gascs. Thesc ~ are from earlierwork at Membrane Tr ' ~cg~ and Research, as already reported in U.S. Patent 4,963,165, since we wcre not able to make ~.. eL~ ~ with pwe l.~l.og~,.. sulfide.
Ex~mple l l. Polyamide-polyether ...~ l,. Pwe gas data A ' ' ~ _. co...~it~ was prepared by coating a polysulfone support --~.nl,. allc first with a thin high-fluY, sealing layer, then with a I wt% solution of Pebax grade 4011 in i-butanol. The ~..~,..~I,. a.l~, was tested with pure gases at a l~ i of 20C and a feed pressure of 448 kPa t50 psig).
15 The results are shown in Table 2.
Example 12. POlya ll;~c polyether Il~ blallcs. Pwe gas data A second Ill~ was prepared using the same materials and tçrhn;que as in Example 11. The ~sults of pure gas tests with this me nbranc are also shown in Table 2. There is good a~o~.ncnl between the sets of results from F.~ 1P~ 1 1 and 12.

r~ lllcdtiol~,. lics of Pebax 401 1 C~ e M
Toeted with Pure Gases ~ ~ . ~ t~
ssYc*c ~at ~ t'~

ll 448 (50psig) 1,6S0 219 11.9 139 18 12 448 t50psig) 1,750 185 9.19 190 20 E~camples 13-18 show the ~ - ru- .--~ of polyamide-pol~ CS exposed to gas ,..;~lu-~,~ under a variety of " ~ ~
r '~ 13 A co.. pos;lt ~ ~ was prepared by coating a layer of a poly ' pol~ ,r ~ 2~ 7~3~7 WO 95111738 Pcr/uss4ll2o99 2s (Pebax grade 4011) onto a polyvinylidene fluoride (PVDF) support 111~ C using the same general tccl~ qu~C as in Example 11. The ,..e~ e was testcd with a two-co~ ~ a gas mixture CV~IA;~ g 4 vol% carbon dioxide, 96 vol% methane at three different feed p.~,;.sw~,;,. 2,809 kPa (392 psig), 4,166 kPa (589 psig) and 6,724 kPa (960 psig). In all cases the p- ~.r ~ ~ side of the IIIC Ill~lanC was at, 5 orcloseto,at..~o~l~h~ icpressureandthemembranewasatroomt,.--~.~lu ~(23C). Thepe . r~;o~
results are listed in Table 3.
Example 14 The same type of,.~ as in Example 13 was prepared and tested with a two-c;~ l gas mixture conci~ing of 970 ppm l,~J.u~,.. sulfide, 99.9 vol% methane at three different feed p.~ ur~;,;
10 2,7791;Pa (388 psig), 4,159 kPa (588 psig) and 6,793 kPa (970 psig). In all cases the ,~.,.. eale side of the ,.. ~,.. l,.. c Wds at, or close to, l" , '- ;., pressure and the Ill~,lllbl - was at room t~ .at (23C). The p~....e..~ion results are listed in Table 3.
E.~.u..~,le 15 The same type of ....,..ll,l - as in r . ~ 13 was Pl~ ~ and tcsted with a three-c~ c ~1 1~ gæ mixture ~ .g of 870 ppm Ly~Lu~.l sulfide,4.12 vol% carbon dioxide and 95.79 vol% methane at three dirferent feed pl~;~aW~ . 2,776 kPa (386 psig), 4,166 kPa (589 psig) and 6,821 kPa (974 psig).
In all cæes the ~ side ofthe membranewas at, orclose to, ~ o~l~k~ ;c pl~ W't and the membrane was at room temperature (23C). Thc p.... ~ results are listed in Table 3.
EYample 16 The same type of membrane as in Example 13 was prepared and tested with a three-c~
gas mixture cc..~ ;..P of 0.986 vol% Ly~ùgc.l sulfide, 4.12 vol% carbon dioxide and 94.90 vol%
met~ - at three different fced pf~a:~w~ . 2,786 kPa (389 psig), 4,145 kPa (586 psig) and 6,800 kPa (971 psig) In all cases thc p. . side of lhe membranc was at, ûr close to, a~ he-iC pl~"~aul~ and the Illc llblanc was at room tempcrature (23C). The permeation results are listed in Table 3.
25 E~ JIC 17 The same type of membrane as in Example 13 was prepared and tested with a three~gas mixture c~ p of 1.83 vol%~.~hu~ sulfide,10.8 vol% carbon dioxide and 87.34 vol% methane at a feed ~.,.,;, .u,~ of 6,759 kPa (965 psig). The ~ ~ side of the mcmbrane was at, or close to, a~ os~h~, ic ~"e;,~w~i and the mcmbrane was at room tcmpcraturc (23C). The ~ ..~c ~l;o~ results are 30 listed in Table 3.
Example 18 Thc same typc of membrane as in Example 13 was l"~,"a,~,d and tcsted with a three-co~ o WO95/11738 ~ 7'13~7 26 PCT/US94/12099 gas mixture ~ w~ .g or950 ppm l"~d,u~;~". sulfide, 8.14 vol% carbon dioxide and 91.77 vol% methane at three different feed p,~~ s. 2,800 kPa (391 psig), 4,138 kPa (585 psig) and 6,793 kPa (970 psig).
In all cases the p - side of the .~ was at, or close to, .~ h. ;c pressure and the ",~,mb, was at room t~ t~ci (23C). The ~, ~ results are listed in Table 3.

p~. ".~ ~;n.~ ~lù~lli~,s of a Pebax~D 4011 C~ c Me"~
with Various Feed Gas G-~ at Three Feed F~l~ia~

2,809 (392psig) - 31 1.9 - 17 13 4,166 (589psig) - 30 1.9 16 6,724 (960psig) - 29 2.0 - 15 1~ 2,779 (388psig) 91 - 1.8 51 14 4,159 (588psig) 74 - 1.8 41 6,793 (970psig) 73 - 1.8 41 2,766 (386 psig) 140 31 1.9 70 16 4,166 (589psig) 115 30 2.0 56 15 6,821 (974psig) 110 29 2.2 52 14 2,786 (389psig) 113 32 2.0 55 16 4,145 (586psig) 103 31 2.0 51 15 6,800 (971psig) 97 29 2.0 48 14 17 6,759 (965psig) 121 34 2.4 50 14 2,800 (391 psig) 93 26 1.6 S8 16 18 4,138 (585psig) 108 32 2.0 52 15 6,793 (970psig) 93 28 1.9 48 14 The following observatiûns can be made ~om the data of Examples 13-18:
1. The pres~ce of carbon dioxide in the feed gæ appears to increæe the fluxes of both L
sulfide and methane through thc ~-.- -l ..- e. For e . ' , a c~ r of the results Of r , 17 14, in which the feed mixture did not contain any carbon dioxide, with those of Examples 15-18, shows that the ~7~3~7 ~vo 95/11738 PCT/US94/12099 hydrogen sul~ide fllrYes are about 25% lowerand the methane fluxes are about 15% lower in F.Y~mpl~ 14.
The i.,c.eas~ fiuY may be due to swelling of the membrane by d;s501~d carbon dioxide.
2 In general, the pressure-nu,..-ali~d fluxes of hyJIu~ sulfide and carbon dioxide decrease with i"~ ,g feed pressure, whereas those of methane increase. The dc~ ,asc in the Lyd~vg.,.l sulfide 5 and carbon dioxide fluxes may be due to c4~ sorption, which results in a lower solubility c~. ~ : .~ (the rat;o of cowentration in the polymer to partial pressure) for cach u~ At the same time, the polymer swells, .~.lLil.g in a higher ~lirr~;v;l~ for all c4l.l~..~,.lts,; c~ .g methane. The net result is an increase in the methane flux and a decrease in the fluxes of the acid gases (Ly~h~.,~ sulfide and carbon dioxide).
3. The hydrogen sulfidel nctll&-c scl~li~ity for three-co ~ r ~ s varies from a low of 48 to a high or70~ although all orthe nc&~u..,.ll~.lb were made at fairly high feed p l~ UI.,S The carbon dioxide/methane scle~livily, also at high pressure, is about 14-16.
Example19 Gasstreamscr,.~ watervapor The e..~.i..~.-t~ of r , ' 15 were repeated using feed gas streams saturated with water vapor 15 by bubbling the feed gas through a water l~ o;-. The .,~ .lt~ were carried out at feed pl~
of 2,772 I;Pa (387 psig), 4,159 kPa (588 psig) and 6,793 (970 psig). The permeate side of the membrane was at, or close to, sî~.. o,~l ;c pressure and the membrane was at room temperature (23C). The ~e.,l,ealion results are listed in Table 4.

Pc ~e ~l;o~ P~upc.li.,s of Pebax 4011 Co -~ it~ Membranes Tested with Watcr-Saturated Gas Mixtures æ v ,~i~x-cc,~ c, 2,772 (387psig) 77.0 18.9 1.03 74.9 18.4 4,159 (588psig) 73.5 20.1 1.20 61.4 16.9 6,793 (970psig) 68.6 18.1 1.17 58.8 15.5 COl.. ~.a.;.lg these results with those of Table 3, it can be seen that the fluxes are aJ~ ly lower (about 40-45% lower) than ~ose cbt ~d in the absence of water vapor. Neither the h~hu~
,~lrd~hll~ northe carbondiu~Jn~ ne ~I~LvitLcs, however, change s;g"-rl~,antly. Fu- ~ .lllUlG, when the lll.,.llbla.l~,s were retesteds with a dry gas strearn, the fluxes returned to the original values.

WOgS/11738 ~ 7~3~ 7 Pcr/uss4ll2 E~xsmples 20-25 show typical c~ .t~, c~ c used to prepare a wne dia~ These s, and others of the sarne type, were used to prepare the wne diagram of Figure I, which shows feed gas carbon dioxide Cl~nC~ iOn on one axis and l,~Lu~n sulfide C;OnC~ on the other. The S diagram was prepared by running a series of ~ sep.u dtiull C~'p~!t~ `;".~ C for l~ulL~ al three~ .- .1 (rn~tL~ ~, carbon dioxide, L~Lug~,n sulfide) gas streams of pal~i~,ul~ flow rates and c~ ;. .nc at a feed pressure of 6,897 kPa ( 1,000 psia). In all cases, thc target was to bring the stream to a pipeline s~ifir~ti~n of 4 ppm hydrogen sulfide and 2% carbon clioxide. Thc ln~ b~ - plop~ .s were accl~m~d to be as follows:
10 More CO2-scle~live ...,..-1.- ~:
Carbon dio.~id~ ,lhane s~ ivily: 20 IIyJ.ug~,.. sulfide/methane sel~liv;ly. 25 Me~ .c flux: 7.5 x 10~ cm3(STP)/~.n7-s .,.... Hg lS More H2s-acle~ meb.~lç.
Carbon liuAidc/melhane ,~.l~liviLy. 13 IIyJIog~,,l sulfide/methane scl~livily; 50 Methane flux: 7.5 x l O~ cm3(STP)/~.ll,7 s .. ,.Hg In cach casc, the methane loss into the ~ s1ream that would occur if a one-stage lll.,llll,l on process were to be carried out was c ~ , and was used to defune zones of least methane loss.
Example 20.
A co~ uh ~ - was carried out for a feed stream of co.. l-o~ 200 ppm l~d~u~-sulfide, l5 vol% carbon dioxide, the ,~ ' mtth- lt Four s:-""~ onC were performed: (i) using only the more hydrogen-sulfide-selc~ , membrane (membrane A), (ii) using only the more carbon-dioxide-sele~ ,-llI,-~-e (111~ Ll - B), (iii) using a: ' . of the more L~J~Ugen sulr~dc-scl~lh,~, n ~ .t. followed by the more carbon~' ' selective .l~l~l.- - (A +B), and tiV) using a 30 of the more carbon~lio~ide sele.,li~_ ~ - followed by the more L~d~ug~,ll-~ulrldc.-selccli~
.l,~,.,L.~e (B + A). The results are listed in Table 5.

2~7~347 Mcml,. - type Mc,.nb. n area M ' - loss Residue H2S Residue CO2 (m2) (%) conc. (ppm) conc. (vol%) A 203 18.1 <0.1 2.0 B 204 18.6 4 0.6 A + B 160 14.3 4 2.0 B + A 160 14.3 4 2.0 The ...~,lh .e losses in all cases are high, because the process design was kept to a simple one^
stage design for COIIIIJ~;SOII P~I~OSCS. The goal of this C~IC~1 was not to design a fully op~
10 process, but to d~ r, which of the possible ~ blallc types would be preferred. It is apparent from the table that a CQ-.~ n~ of the two membrane types would be indicated for treating a stream of this ~IIl~os-li~
E.~...ple 21.
A c~ .~ . ' ' was performed as in F , ~~ 20, using a foed stream of co- . .l~u~ 70 15 ppm l-~u~w- sulfide, 10 vol% carbon dioxide, the ll....-;..~lr. " ~ The results are listed in Table 6.

Mc...l,. - type Mc...l..~lc area M~ loss Residue H2S Residue CO2 (m2) (%) conc. (ppm)conc. (vol%) A 171 14.7 <0.1 2.0 B 161 14.0 4 1.0 A + B 130 11.5 4 2.0 B + A 130 11.5 4 2.0 E~ ,'e 22.
By ~ scts of c~c ' - as shown in r .1 21 and 22, the b ~ ~ line betwoen zones C and D, based on the stated ~ as to .. ~,.. l.~ ~ p~ ~- ~~ and operating co~ "~, was ~ n~ to be as follows:

WO 95tll738 PCTIUS94/12099 CO2 content of feed gas H2S content of feed gas (vol/O) (ppm) F,-,amFI~ 23.
Cr~ ;ri~C similar to those d~ bc~ in F~ .plr 22 were carried out to ~ ...;.e the position of the bou..d...y line between zones D and B. The position of the bOu~da~y was d t~",.;--~ to be as follows:
CO2 content of feed gas H2S content of feed gas (vol%) (ppm) 13,000 120,000 Example 24.
The ~nc of Example 23 were repeated to show the effect of higher or lower S.,l~liv;L~ on the zone bou,ld......... ics. Representative c~ ;o~c were p~ d ~ .g a Ly~LOg~.l sulfide/methane selc.livily of the more l~u~ sulfide-selective Ill~ lC of 30, 40 or S0, and a carbon ~' '/,.~ll.. ,.r_s.l~livi~yofthemorehydrogen-a~lridc-s_le.li~,~.-"~.,-l,- -of10,13or15.Theresults are plotted graphically in Figures 9 and 10. As can be seen, the Zonc B/D bu~d~uy moves to the right as thc ability of the lll~l~l to separate carbon dioxide illlplU . _5. Likewise, the bOulllia.y moves to the right as the sel~LiviLy for hydrogen sulfide over methane dc~,- ases. Although the area where the more Lydrog~ ~I-sulfide-s~ membranes should be used is larger at lower l,~dlug~n sulfidc~ .Ll,d~,c 3S selc~ liv;Ly, the ."~ tl,~c losses ç-~ d in using the .". ",I,.~c will be greater.

F~ le 25.
The, 1 - of F~ F'e 23 were repeated - ~ .g different values for the feed pressure.
Rep~s.,llative c~ c were ~.Çu---.ed ~ ~---.-;-.g a feed pressure of 5,517, 6,897, or 8,276 kPa ~, 21~3~7 (800, 1,000 or 1,200 psia). The results are plotted graphically in Figure 11. As can be seen, the zone b~ daly is l~,lati~.,ly l~s.,.~ ,_ to changes in the feed pressure.

F,x~ ,les 26-29 show .~ ,s7e,~ _ yluUiS~,S using the more l,~Lugen-sulfide-sclecli~_ S ",e",~,a lc only.
le 26 A very simple o"c-st..b_ Il..,.llll ~ process was d~ fd to handle a gas stream c-..a~ .g 100 ppm hydrogen sulfide, 0.1 vol% water vapor, 4 vol% carbon dioxide and the ~ c~ ' e, at a feed pressure of 6,897 kPa (1,000 psia). A basic s~ of the process is shown in Figure 2, where 10 numeral I ;..~ ., the bank of Illc.~lblallc ~l~ C, and the feed, residue and yc ...~ ~ç streams are in-lic~ted by numerals 2, 3 and 4 r~ y. The process was ~Ccum~d to usc one bank of more l,.y.l.u~.,-sulfide-scle~ , mc.l~ cs having the following dlala~,t~,~i.7lic~.
Hydrogen sulfide/l"~,lL~.c sclc~ti~ly: 80 Water v .~I/lll~,lhallc scl~li~ity: 1,000 15 Carbon diOxidc~5,-cthâ..c scl~livily; 12 Me~hane flux: 1 x 10~ cm3(STP)/cn,?-s ~,.--Hg The co...posiliu-,s and flow rates of the pr.,..~ , and residue streams were c~lc--' ~ and are 20given in Table 7.

STREAM FEED RESIDUE PERMEATE
Flowrate(Nm3/min) 28.3 (l,OOOscfm) 25.6 (903scfin) 2.7 (97scfm) CH4conc. (vol/O) 95.9 98.1 75.6 C02conc. (vol/O) 4.0 1.9 23.2 H2S conc. (ppm) 100 4 995 Watervaporconc. (vol/O) 0.1 2ppm 1.0 The Ill~,.lll,l ,c area used to perform such a ~p~ was calculated to be about 70 m2. The stage cut was just under 10% and the ' ~ loss into the p~ was 7.5%. The process yluduecs a residue stream that simultaneously meets pipeline specification for carbon dioxide, l,~dlùgc.~ sulrlde and water vapor. The low grade permeate gas could be sent to the foul gas line.
F.Y~mp~e 27 3S The simple design of Example 26 is only possible for certain cases where the raw stream to be ~17~3~7 WO 95/11738 Pcrruss4/l2o99 treated contains an .. , ~.o~ .;ate balance of l-~J-u~ sulfide and carbon dioxide. In many cases, a more C- r ~ U~ 1 design is needed to improve the methane ~~ and moet pipeline Spf'~ ;~';C '~ ;U-~C
without u.~ g A process was designed to handle a 28.3 Nm3/min (l,OOO scfm) gas stream c~ .;..g l,OOO ppm hydrogen sulfide, 0.1 vol% water vapor and the ~ , so as to koep methane loss in the stream below 2%. The process uses a two-stage ....,.,L.~c S~ UUtiOIl system in which the pc.~lc~le rrom the first bank of Ill~nlbl - modules b~m~C the foed for the second bank. A basic S~ G~ ;Ç of the process is shown in Figure 5, where numeral 10 ;~ es the first stage bank of l~f ~ f modules and numeral 18 indicates the second stage bank of ...~....~-al,e mr~ lf c The ;..~ --g gas stream 9 is at 6,897 kPa (1,000 psia) and is mixed with the residue stream 20 from the second stag to form the reed gas stream 21 to the first ",cn~ , stage. The pe",.cdle stream 12 from the first stage is l~lll~l~i to 6,897 kPa (1,000 psia) in c~ ,;,aor 13. The co"".,~,~s~ stream 14 passes to chiller lS, wherewatervapor is cn.ul~ andwateris removed as liquid stream 16. The non-co...~ c~ stream 17 enters the second ",..."b, - stage 18, where further s~u of hydrogen sulfide takes place. The lS residue stream from this stage is .~uculat~ within the process. Both --e"-l,-.u-c stages were ~cc lmed to use more hydrogen-sulfide-~el~Li~ ~. ",~ Inl)l ulCS having the following ~ I.~ ~.,t., i:,Li~ ..
Hydrogen sulfidc~ .ll-.u,e scl~livit~; 50 Water v, ~"~,I}.~.c sel~livity: 1,000 MelL~c flux: 7.5 x 10~ cm3(STP)/c---7-s-.,.. -Hg The cc~ ;ol~s arld nOw rates ofthe first and second stage p~ f~Vtl and residue streams were c~ and are given in Table 8.

~'o 95/11738 2 17 4 34 7 PCT/US9.1/12099 STREAM ¦ FEED ¦ RESIDUE ¦ PERMEATE
FIRST STAGE
Flowrate (Nm3/min) 34.0 (1,200sefm) 27.9 (985scfm) 6.1 (21~sefm) CH4eone. (vol%) 99.82 99.99 98.99 Watervaporconc. (vol%) 0.08 0.0 0.45 H2Seonc. (vol%) 0.10 4ppm 0.55 SECOND STAGE
Flowrate (Nm3/min) 26.1 (215scfin) 5.7 (202scfn) 0.4 (13scfm) CH4conc. (vol%) 99.42 99.89 92.12 Watervaporeonc. (ppm) 330 21 5,015 H2Seonc. (vol%) 0.55 0.1 7.4 The ~ llbl - area used to perform such a separation was e~ to be about 280 m2 total, lS 265 m2 in the first stage and 15 m2 in the second shge. The residue stream 11 from the first stage meets pipeline ~ fi~a~ nc The ~ - stream 19 from the seeond shge eonhins a high enoughco"cc.l~lion of L~ u~l sulfde to be passed to a Claus plant for sulfur ~ .y unit, or to a liquid redox process, such as LO-CAT, Sulferox, II~ or Stretford. The overall .~ hanc loss into the seeond stage pe.,-.eale is very low, at just about 1%.
20 E~;ample 28 A process was designed to handle a 28.3 Nm3/min (1,000 scfin) gas stream c - ling 1,000 ppm Lydlog.,.l sulfide, 4 vol% earbon dioxide and the ,~ The process uses a two-stage lllbl - s~ u.,tio.. system in which the ,~ ~ from the first banl; of membrane modules ~4~ -- s the feed for the second bar~ The proeess s- h -- ~t;c iS as shown in Figure 5, exeept that no c~ - 15 25 is used. Numeral 10 i- ' - the first shge bank of mambrane modules and numaal 18 ' - the second stage bank of membrane modules. The ~g gas stream 9 is at 6,897 kPa (1,000 psia) and is mixed with the residue stream 20 from the second stage to form the feed gas stream 21 to the first membrane stage. In this ease, the ~ ~ ~ stream 12 from the first stage is ~ aa~d to 6,897 kPa (1,000 psia) in co -~ ~aor 13, than passed without any condensation taking place as cO~ as~l stream 17 to the second "-~,.. ,1" - stage 18, where further a~ L ' of h~ ,5,e.l sulfide hkes place. The residue stream from this stage is .~' ' within the process. Both membrane shges wae æCcllm~
to use more Lydlugcn-sulfidc-sel~~ LI~.. Ics having the following ~ ic~i~c-~ ~ 74 ~ 4 7 34 Pcr/uss4/l2o99 ~

Hydrogen sulfide/methane scl~liv;l~: 50 Carbon dioxidw~"~ll,~lc scl~livily. 13 Methane flux: 7.5 x 10~ cm3(STP)/. ",~-s ~n,Hg The ~ ~ r~ ofthe ~rst and second stage permeate and residue streams were calculated and are given in Table 9.

STREAM ¦ FEED ¦ RESIDUE ~ PERMEATE
FIRST STAGE
Flowrate (Nm3/min) 34.5 (1,220scfm) 27.3 (964scfm) 7.2 (256scfm) CH4conc. (vol%) 93.0 98.86 71.8 CO2 conc. (vol/O) 6.9 1.14 28.3 H2S conc. (ppm) 1,000 4 4,733 SECOND STAGE
Flowrate (Nm3/nun) 7.2 (256scfm) 6.2 (220scfrn) 1.0 (36scfrn) CH4conc. (vol/0) 71.8 80.0 19.6 CO2conc. (vol%) 28.3 19.9 77.7 H2Sconc. (vol%) 0.47 0.1 2.7 The ",c."~-~.e area used to perform such a separation was ePlcl~lPtP~d to be about 244 m2 total, 232 m2 in the first stage and 12 m2 in the second stage. The residue stream 11 from the first stage meets pipeline ~I.~;r.~r ;OI.c The ~ t., stream 19 from the second stage contains a high enough 2S ~v,,~lL dtiu.. of hydrogen sulfide to be passed to a Claus plant for sulfur ,~ unit, or to a liquid redox process, such as LO-CAT, SuLferox, II~ or Stretford. The overall methane loss into the second stage pr.. .--. . ~' iS very low, at about 0.7%.
Example 29 The c~ ,.c Of F . '~ 28 were repeated wilh a 28.3 Nm3/min (1,000 scfm) gas stream ~ g 10,000 ppm hydrogen sulfide, 4 vol% carbon dioxide and the ,. -.. ~ The results are given in Table 10.

~o 95111738 2 1 7 ~ ~ ~ 7 PCT/US94/120ss STREAM ¦ FEED ¦ RESIDUE ¦ PERMEATE
FIRST STAGE
Flowrate (Nm3/min) 37.7 (1,330scfm) 26.9 (950scfm) 10.8 (380scfm) CH4conc. (vol/O) 91.0 99.4 700 CO2conc. (vol/0) 8.0 0.6 26.5 H2Sconc. (ppm) 10,000 4 3.5vol%
SECOND STAGE
Flowrate (Nm31min) 10.8 (380scfm)9.4 (330scfm) 1.4 (50scfin) CH~conc. (vol%) 70.0 78.6 16.2 CO2conc. (vol/0) 26.5 20.4 64.6 H2Sconc. (vol%) 3.5 1.0 19.2 The ~ bl.ulC area used to perform such a S~i~aldtiOU was c '--_' ' to be about 353 m2 total, 339 m2 in the first stage and 14 m2 in the second stage. The residue stream 11 from the first stage meets pipeline speeifi- ' ~nC The ~""...~t~, stream 19 from the second stage contains a very high L~
sulfide coll~llt-ation. The methane loss is less than 1%.
r , ~ 27-29 illustrate the benefits of two-stage pl~;,~S in both l~lu~ g m~ot~ loss and raising the hydrogen sulfide concentration of the waste stream. In rY~ npl 27-29, thc feed &0 both raw and sfter mixing with recycle stream 20, is in zone B.
Example 30 A process was designed to handle a 28.3 Nm3/min (1,000 scfm) gas stresm &~ g 1,000 ppm L~l~ l sulfide, the .~ methane. Thc process uses a membrane separation system as shown in Figure 8. N~ll~,.dls 38, 44 and 47 indicate the three banks of membrane ~ , all contain the more Lydlog~.~-sulrldc-s~ membrane. The ;--~- -: g gas stream 36 is at 6,897 kPa (1,000 psia) snd is mixed with the residue stream 49 from module(s) 47 to form the feed gas stream 37 to the first Ill~ 7 stage. The~-.--- --t~,strearn,40,fromthefirststageis1~lll~ ~in &01ll~ .0l 42. Co...~ sol 42 drives two - ...l.~....~ units, the second stage unit, 44, snd an auxiliary module or set of m~ll.oc, 47, that sre c~)---l-,,t~d on the p- .. ~'~ side either directly or uldu~lly to the inlet side of the C~ O~, SO as 30 to form a loop. Thus, pcrmeate stream 48 may be merged with pl - stresm 40 to form c~....l.:- ~
stream 41. The l~lll,l~d, r ' ' stream, 43, passes as feed to membrane unit 44, and the residue stream,46, from .. -~ unit 44 passes as feed to membrane unit 47. Pe.lllcdt~ is withdrawn from the W O 95/11738 ~ ~ 7 ~ ~ ~ 7 PC~rrUS94/12099 loop as stream 45 and the treated residue exits as stream 39. This system configuration is particularly useful in ~ J~ where the l-~Log~,.. sulfide content of the raw stream is .~,lali~,~.ly low, yet flaring is not an option and the stream has to be brought up to a viable c~ucc-d,-,lion for sulfur I~U~ . A series of calculations was carried out by keeping the area of membrane unit 38 con~t~nt but varying the relative S areas of ~ . units 44 and 47. The char~ ,c of the mh~-~.c were a~lm~l to be as in . '- 28. The results of the c~ ' - are given in Table 11.

Membrane Area (m2) P.,.".. ,ale conc.
(vol/O) Unit 38 Unit 44 Unit 47 Total 10242 0 18 260 2.65 242 10 11 263 4.26 242 15 8 265 5.77 242 20 6 268 8.92 242 35 2 279 19.7 1~242 50 0.4 292.4 55.0 The residue stream 39 from the first stage meets pipeline -r ~' ' S. A high C~llc~ on of L~r~ugc--sulfide in the waste pe, - strearn can be a~ ~I with an 1, ~ u~ choice of ",~,."~, ~ areas.
This type of design could also be used in - - where ~ of the two ",.,.. ,~, -types are i~ ;c ~t~

E~camples 31-34 deal with streams in which the feed u~ n~;l;o~ is in zone D, so that a c~- k~ ;ml of Ill.,..LI .c ~pes is ' ~
E~_.",lc 31 2~ A process was d ~ rd to handle a 28.3 Nm3/min (1,000 scfm) gas stream c~ -.g 60 ppm h~v~;w~ sulfide, 15 vol% carbon dioxide and the l~ f - methane, a co..-~ ;o~ that falls in Zone D
of Figure 1, but close to the boundary between zones C and D. The process uses a ",.",L, sep~udlivn system as shown in Figure 7. Numerals 23, 26 and 32 indicate the three banlcs of Ill~l~la~ .~.h.l..l..~" 23 eontains the more h~Lugcn j~llr.dc~scl~li~ lc, 26 and 32 eontain the more earbon-dio2cide-sclecli~ m~ b~ c. The ;---~ g gas stream 22 is at 6,897 kPa (1,000 psia) and is mixed with the residue stream 34 ~om the second stage to form the feed gas stream 35 to the f~st ,.~nL, - stage. The residue stream,24, fiom the firstbar~cofmodules passes as feed to the second bank of the first stage, 26.

~tvo 95/11738 2 1 7 ~ 3 ~ 7 PCT/US94/12099 In this case, the p~ te streams 25 and 28 from the two steps of the first stage are cc l.lbl.led as stream 29 to be ~ ,a5c~;l in COIII~ ,aSOI 30, then passed as COIll~ SCd stream 31 to the second lll.,.nl~l le stage 32. It will be a~J~U. ~ to those of ordinary skill in the art that two separate C~IlI})l~aSOla could be used and the stream colllb;llo~ after coln~ a;ùll. Also, in cases where the stream to be treated contains S water vapor, the system could include a co~ c- as in Figure 5 to c~ -cc p~ water vapor.
The c~ .u~ of stream 31 was in Zone C, so that the more carbon~l;o~idc-sel~li~ lll.,.llbl- .IG was chosen for the second stage. The chal iaL;~,a of the two types of lll~..lll,l ~c were ~C~l~nPrl to be as follows:
More ll~.llog~ sulfide-sel~~ , membrane:
10 Hydrogen sulfidG/ ' - scl~:~ivi~y: 50 Carbon d;u.~;dc/lll.,lhane sel~livily. 13 Methane flux: 7.5 x 10~ cm3(STP)/~I.l' 5 ~lllHg 15 More carbon-d;o.~idc-selG~ . me.llbl ~:
Hydrogen sulfidc/methane sel~l;vi~: 25 Carbon dioxide/methane selc.,livily. 20 Mclhane flux: 7.5 x 10~ cm3(STP)~clll' 5 .,lllHg The c~ of the various streams were c~lr~ and are given in Table 12.

Stream# CH~conc. (vol%) H2Sconc. (ppm) CO2 conc. (vol/O) 22 85.060 15.0 84.060 16.0 24 90.210 9.8 27 98.0 1 2.0 36.1456 63.9 28 52.351 47.7 31 45.5223 54.5 33 7.5 407 92.5 34 78.660 21.4 The membrane areas rcquired were as follows: 66 m2 for lll~.llbl - 23, 120 m2 for membrane 26 and 22 WO 95111738 2 ~ ~ ~ 3 ~ 7 Pcr/uss4/l2099 m2 for ~I.c.~lbranc 32. The residue stream 27 from the first stage meets pipeline srerif~: - The pc.,lledt., stream 33 from the second stage contains about 400 ppm Ly~ugcn sulfide and the overall methane loss is about 1%.
Fxamrle 32 S A process was dcsigned to handle a 28.3 Nm3/min (1,000 scfm) gas stream c~ .;.. g 200 ppm L~r~hu~;.,n sul de,15 vol% carbon dioxide and the ,~ f, mt th~nt, a c~ Q~ that falls in Zone D
of Figure 1. The process uses a - ~ s r - system as shown in Figure 7. Numerals 23, 26 nnd 32 indicate the three banks of membrane n ~ ' ' , 23 and 32 contain the more L~ ugc.l-sulfide-selc~
...~ .~1.,,..~t,26 contains the more carbon~ selective m~ bl~IC. The i..r~ gas stream 22 is at 6,897 kPa (1,000 psia) and is rnixed wilh the rcsidue stream 34 from the second stage to form the feed gas stream 35 to the first ~ bl~IC stage. The residue strcam, 24, from the first bank of modules passes as feed to the second bank of the first stage, 26. As in F ~ le 31, the ~ ""~ streams 25 and 28 could be c-,...b:..r~ before or after l~lll~ a;UII, and a cQ.~tl,r..~r.r to rernove water vapor could be inrl~ltit~i The ~,Lal...,l~,i ,li.,s of the two types of Ill~.lllbl~lC were ~sl~mt~i to be as follows:
15 More lly~u~ l-sulfidc-sclccli~ IC.IIIbl Hydrogen sulfide/methane scl~liv;ly: 50 Carbon dioxidc/ ' - scle~,livil~. -13 Methane ilux: 7.5 x 10~ cm3(STP)/~,Ill7-s ~,mHg More carbon~ioxide-sclc~ _ membrane:
Hydrogen sulfideJh~clhL~c scl~liv;ly. 25 Carbon dioxidc,'~ ' ~ sclec livily; 20 2S M~ 1ux: 7.5 x lû~ cm3(STP)/~ s .,I"~Ig The ~Ill~iliol~ of the various streams were 1~ 'a I and are given in Table 13.

~o 9S/11738 2 ~ 7 ~ 3 4 7 PCr/uss4/12099 Stream# CH4conc. (vol/0)H2Sconc. (ppm) CO2 conc. (vol%) 22 85.0 200 15.0 64.0 200 36.0 24 68.0 130 32.0 27 98.0 4 2.0 13.2 1,443 86.7 28 26.7 294 73.3 31 25.2 427 74.8 33 3.0 1,447 97.0 34 29.9 200 70.1 The ~ lJ-a-~ areas required uere as follows: 21 m2 for ~ bl~ulC 23, 248 m2 for 1ll~.. lll.l. IC 26 and 17 m2 for ~ 32. The residue stream 27 from the f~rst stage meets pipeline sp~ific~tionc The permeate stream 33 from the second stage contains about 1,500 ppm Lyd~o~,.l sulfide and the overall methane loss is about 0.4%. The feed strcam to the second stage bank of ~ 32, contains 427 ppm hydrogen sulfide and 75 vol% carbon dioxide, a C4".1~C ~ that falls in the more carbon-dioxide-scle.,li~ --e.ll~l~lc zone of the zone diagram. IIo.._~w, since it is not required to meet pipeline ~ifi~ n for the residue stream from the second stage, an v~ ; f~ design plU . ;~S better h~L`~.,.-sulfide .~,co~c.y if the more L~Lu~ -sulfldc-3cl~li~ membrane is used.
Example 33 A process was dcsigned tohandle a 28.3 Nm3/min (1,000 scfm) gas strcam co-~ g 1,000 ppm hydrogen sulide, 15 vol% carbon dioxide and thc ~--~ f ~ me h~n.~:, a c ~.n~ that falls in Zone D
of Figure I, but close to the boundaly of Zone B. The process uses a membrane separaffon system as showninFigure7. N~ als23,26and32indicatethethreebanksoflll~ lal~C ~ 23and32 contain the more hy.hug.,.l~ lrlde-sdc~,L~,~, membrane; 26 contains the more carbon-Lo,~ide-scl~li~, -..h~ f- The incoming gas s1ream 22 is at 6,8971~a tl,000 psia) and is mixed with the residue stream 34 from the second stage to fo~m the feed gas strcam 35 to the first lll~ c stage. The residue stream, 24, ~om the fL~t bank of modules passes as feed to the second bank of thc first stage, 26. As in F.~
31 and 32, the p~ ~ streams 25 and 28 could be CC"~ -'f'd before or after ~ 5;0ll, and a co~d~ -~Cf ~ to remove water vapor could be inCl'l(lf~ The characteristics of the two types of membrane WO gS/11738 Pcrluss4ll2o99 ~7~3~ 40 were accllmpd to be as follows:
MoreLy~Lug~ sulfide-sel~li~,~ membrane:
Hydrogen sulfide/methane scl~livity: 50 Carbon ~Lu~idc~ thane SeIG~ Y;lY: 13 Methane fiux: 7.5 x 10~ cm3(STP)/~.".~-s ~."lHg More carbon-dioxide-scl~li~_ ".~ .n~, 10 Hydrogen sulfide/, ' - sel~livity; 25 Carbon dioxide/methane sel~livity. 20 Methane flux: 7.5 x 10~ cm3(STP)/~ s ~.. Hg The c~.. po~ nC of the various streams were c~lcll~ and are given in Table 14.

Stream ~ CH4 conc. (vol%) H2S conc. (ppm) CO2 conc. (vol/O) 22 84.9 1,000 15.0 63.9 1,000 36.0 24 79.7 70 20.3 27 98.0 4 2.0 17.0 3,770 82.7 28 37.4 221 62.6 31 26.6 2,084 73.2 33 3.1 7,390 96.2 34 31.5 1,000 68.4 The ~ areas required were as follows: 119 m2 for membrane 23, 188 m2 for In~ IIILI~G
26 and 17 m2 formembrane 32. The residue stream 27 from the fust stage meets pipeline ,~ u~c 30 Thc pc~mcate stream 33 from the second stage contains about 0.7 vol% Ly~lluL_.~ sulfide and the overall methane loss is about 0.4%.
As with FY~ e 31, an c")t; -;-- ~ design uses the more 1..~ l,og~ .~-a~lfide-scl~lhr~ 1. IC for the second stage.
F.Y~mple 34 3S A process was desigld to handle a 28.3 Nm3/min (1,000 scfm) gas stream c ~ L 100 ppm h,.h~ ., sulfide, 4 vol% carbon dioxide and the ,~ mPth~nP~ a c>~po~i~;o-~ that falls in Zone B

2~ ~3~7 !vo 95/11738 PCT/US94/12099 of Figure 1, but so close to the buul~Ly of Zone D that the c .. . ~ is just within Zone D after mixing with the recycle stream from the second membrane stage. The process uses a -~ e S~aldtiOn system as shown in Figure 7. Nurnerals 23,26 and 32 indicate the three banks of ,-.w.-b, ~ e~, 23 and 32 contain the more hydlogcn sulrldc-s~ _ membrane; 26 contains the more carbon-dioxide-sel~~
S .. ,w-L, - The incoming gas stream 22 is at 6,897 kPa (1,000 psia) and is mixed with the residue stream 34 from the second stage to form 1he feed gas stream 35 to the first ..,~,..ll,.- nc stage. The residue stream, 24, from the first bar~c of mod~es passes as feed to the second banlc of the first stage, 26. As in Examples 31-33, the permeate strearns 25 and 28 could be or~ before or after 1~4111~n~ S;O11, and a CC~
to remove water vapor could be included. The r~ h - ;~lir~ of the two types of lll~nll,l."le were ~sllm to be as follows:
More hydrogen-sulfidc-scl~~ IIIC--IblallC.
Hydrogen sulfide/methane sel~livily; 50 Carbon dioxide/~ l..u.e sel~liviLy; 13 Methane flux: 7.5 ~ 10~ cm3(STP)/~,.. ? s ~.. Hg More carbon-dioxide-s~,l~liv~, membrane:
IIyL~,ge" sulfidel ~ nr scli~livi~: 25 20 Carbon dioxide/ ' o ;~ iYily; 20 Methane flux: ? 5 x lo~ cm3(STP)/~""~-s .""Hg The c~....l.o~:~;Qn~ of the various streams were l '~l ~ and are given in Table 15.

SLream# CH~conc. (vol/O) H2S conc. ~pm) CO2 conc. (vol/O) 22 96.0 100 4.0 94.0 100 6.0 24 98.0 4 2.0 27 98.1 4 1.9 70.3 741 29.7 28 75.7 57 24.3 31 70.3 737 29.7 33 20.0 3,680 79.6 34 81.2 99 18.8 Wo95/11738 ~ 43~ 7 42 PCr/Uss4/12099 The ~ bla l areas required~vere as follows: 131 m2 for .,~.."1 -_~ 23, 1 m2 for Ill ~ntilanf~ 26 and 9 m2 for ~ "~,auc 32. The residue stream 27 from the first stage mcets pipeline ..l,e~ ;o..c The p ~,r ~f stream 33 from the second stage c ontains about 0.4 vol% l,~ug~,,. sulfide and the overall methane loss is about 0.5%.

"' - 35-38 ~Ulllyl~G the pl~ -- of different types of ..~.IL- - process for various fced gas ~ ~s~ nc The p.u~ ,s are not u~)t;...;~fA but are simply ;ntfn~lf~d to highli~ht the difference between the l ~yf~Li~c perf~
10 Ex~nple 35. No carbon dioxide: ."od~,dte amounts of hy~Lu~_.l sulfide A one-stage "e-,~ u,e. process was ~e~i~fxl to handle a gas stream f~ --p, 100 ppm hydrogen sulfide, the l~ .1. mf th~nf at a feed pressure of 6,897 kPa (1,000 psia). The process is as shown in Figure 2, wvhere numc~al 1 indicates the bank of l~ c m~ f C, and the fced, rcsidue and p. "~ ,-t strearns are ' ~ by ' 2,3 and 4 l~i"~Li~,_ly. The process was ~c~l-mf d 15 to use one bank of more carbon-J;~ .~idc-sclc.,li~_ membranes having th-e following chala~,t-.;~li i...
Hydrogen sulfidcllll. lL ~c sclc~livily; 25 Carbon dioxidc/methane s.,l~Livit~: 20 Me~h~nf 1ux: 7.5xlO~cm3(STP)/clll7-s.,lllHg The cUll-~ u~iLiolls and flow rates of the p~ ...,~Oe and residue strcams were c~ .,d and are given in Table 16.

STREAM FEED RESIDUE PERMEATE
Flowrate (Nm3/1r~in) 28.3 (l,OOOscfm) 23.6 (833scfin) 4.7 (167scfm) CH4conc. (vol%) 99.99 99.99 99.94 H~S conc. (ppm) 100 4 580 The ,llc.l-bl~lc area used to perform such a ~ wæ c~ 'f~ to be about 200 m2. The stage cut was 17% and the methane loss into the p~ ~ was 17% also.
Theprocess design calculation was repeated using more L.~ho~.l-sulfide- ~ membranes having the following characteristics:
II~nLu~ sulfid~ll~,lllu~c scl~;fivi~: 50 35 Carbond;v,~idcJ-- ~ scl~Livity; 13 lo 95/11738 2 ~ 7 4 ~ ~ 3 PCT/US94/12099 Mcll,~e flux: 7.5 x 10~ em3(STP)/c."'-s c",Hg The c~ o~ c and flow rates of the p ~ t~, and residue streams were e~ ^A and aregiven in Table 17.

- STREAM FEED RESIDUE PERMEATE
Flowrate (Nm3/min) 28.3 (l,OOOsefm) 25.3 (B92sefrn) 3.0 (108sefm) CH~eone. (vol/O) 99.99 99.99 99.91 H2Scone. (ppm) 100 4 890 The "-~,."I"a.,e area used to perform sueh a separation was c ~ ~ to be about 130 m2. The - stage eut was 10.8% and the methane loss into the p~ was 10.8 % also.
Co--l,z.".g the two c~ t;.~c the loss of methane into the p~ f~le through the more 15 L~l-ugw~-sulfide-scl~1i~ membrane is about 2/3 of that through the more earbon-d;~,~dc-scl~1i~
."~."I"~,c. The p~.",ca1e stream from the more l,~d,u~j.,.,-sulfidc-sclf~li~ is about 2/3 the volume and 1.5 timoe more eoneentrated than the p ... t~ stream from the more earbon divA,dc-s_lc.,~
llblall~" making further treatment mueh easier. The proeess with the more hyJ~ug~ -sulfidc-3~.1~1i~, ",~,.nl.,. ,c also uses less Ill~,llll,l -..c area.
20 Example 36. Small amounts of earbon dioxide: moderate amounts of h~n sulfid A one-stage Il~ IC proeess was rlf`;~'/'d to handle a gas stream co-~t~ .g 100 ppm hydrogen sulfide, 4 vol% earbon dioxide and the r~ , at a feed pl~ aul~ of 6,897 kPa ( 1,000 psia). The proeess sehematie is as shown in Figure 2, where numeral 1 indieates the bank of membrane mod--~ec, and the feed, residue and permeate streams are indieated by numerals 2, 3 and 4 25 .~,;,~.cc1i~1y. The proeess was ~ med to use one bank of more earbon~;o~idc-scl~1i~, membranes having the following eha~ t~
Hydrogen sulfide/methane sclc~ ,;1y. 25 Carbon~' ' h..~ csel~1ivily. 20 Methane flux: 7.5 x 10~ em3(STP)/~ s ~,Il,Hg The c~ and flow rates of the permeate and residue streams were ealeulated and are given in Table 18.

~ 7~347 wo 95/11738 PCT/US94/12099 STR~AM FEED RESIDUE PERMEATE
Flowrate (Nm3/min) 28.3 (l,OOOscfm) 22.8 (807scfm) 5.5 (193scfm) CH4conc. (vol/O) 95.99 99.72 80.38 S C02conc. (vol/O) 4.0 0.27 19.5 H2S conc. (ppm) 100 4 502 The ..e..~ c area used to perform such a pr a ~ was ~~' ' ~ to be about 200 m2. The stage cut was 19% and the methane loss into the permeate was 16%.
Tl2e process design r~l ,u' was repeated using more Ly~llug~,. sulfidc-sclccli~ ll-b-all~,s having thc following cl-a~cl~";ali~s.
Hydrogen sulfide/methane scl~livity: 50 Carbon dioxide/lncll~anc s ~ liv;ly: 13 IS Methane flux: 7.5 x 10~ cm3(STP)/.. 7-s ~,.. IIg The co.. posiliù. s and flow rates of the p~ t~, and residue streams were ~ and are given in Table 19.

STREAM FEED RESIDOE PERMEATE
Flowrate(Nm3/min) 28.3 (l,OOOscfm) 24.8 (876scfm) 3.5 (124scfm) CH4conc. (vol%) 95.99 98.62 77.34 C02conc. (vol/O) 4.0 1.38 22.58 H2Sconc. (ppm) 100 4 780 The ..~".-b. ~c area used to perform such a separation was calculated to be about 120 m2. The stage cut was 12% and the methane loss into the ~ - was 10%.
Co---y -..g the two calculations, the methane losses, ~ conc."l~-ation, ~ .l~ t~. volurne 30 and ~ ~ area are once again better with the more h~.Lugc,- sulfide-a~ ,lllbl~ulC. It should also be noted that the more carbon~Lu.~.dc-sel~li~,~, membrane, in order to bring the residue stream hydrogen sulfide c-~ hation to 4 ppm, reduces the carbon dioxide co.~ lh~tion to the low level of 0.27 vol%, which means that the gas stream has been s;g '~ ly u~ -uce;~
F ,17 37. Large amounts of carbon dioxide: .-od~ amounts of l.~/d.ugcn sulfide A one-stage ~ ,nl~- ~ process was .l~i~ to handle a gas stream C.~ -- .. g 100 ppm ~ 7~3~7 ~lo 95/11738 PCT/US94/12099 hydrogen sulfide, 30 vol% carbon dioxide and the ,~...A;...~f mfthAne, at a feed yl~,aa~ of 6,897 kPa (1,000 psia). The process Sr~ I;c iS as shown in Figure 2, where numeral 1 ~ the bank of membrane m~-llr~A and the feed, residue and Ih ~ streams are ;..~ ted by llwll~lals 2, 3 wnd 4 r~ayf~~ . The process was A~s~mfYl to use one bank of more carbon~l;o"ide-sel~Li~, membranes S having the following cl,w_ iali~s.
IIydlugc,~ sulfldf~ Lhwle sel~Li~,ity 25 Ca~bon d;o..idcJ~ ' - sel~li~,ity: 2û
Methane flux: 7.5 x 10~ cm3(STP)/~,I.,7-s ~,."IIg The ~ros;l;o~c and fiow rates of the pe.-"e.,t~. and residue streams were c~lrl~ ed and wre given in Table 20.

STREAM FEED RESIDUE PERMEATE
Flowrate (Nm3/min) 28.3 (l,OOOscfm) 16.9 (598scfm) 11.4 (402scfm) CH~conc. (vol%) 69.99 97.99 28.34 C02conc. (vol/O) 30.0 2.0 71.64 H2S conc. (ppm) 100 3 244 The ,.. ~ area used to perform such a s~F - was r~lrlll- ' to be about 150 m2. The stage cut was 40% and the methane loss into the 1~ ; was over 16%.
The process design cLl- "15~ was repeated using more h~.Lug~ sulfidc-sel~~ , mc.nl",...es having the following -' ~. - ;aL;~,a.
II~nllugen sulflde/ ' - - sel~LYity. 50 25 Carbon d:o~ sel~li~,ity: 13 MeLl.anc flux: 7.5 x 10~ cm3(STP)~,.. ?-s-c.. Hg The ~lll~a;l;ons and flow rates of the ~ and residue strearns were c~le~ ed and ~re given in Table 21.

STREAM FEED RESIDUE PERMEATE
Flowrate (Nm3/min) 28.3 (l,OOOscfm) 15.3 (541scfm) 13.0 (459scfm) CH4conc. (vol/O) 69.99 98.0 37.0 C02conc. (vol/O) 30.0 2.0 62.97 H2Sconc. (ppm) 100 4 218 WO 95/11738 2 ~ 7 ~ ~ ~ 7 PCT/US94/12099 The ,l,~.nbl~c area used to perform sueh 8 separation was c~lc~ t~ to be about 240 m2. The stage cut was 46% and the methane loss into the pl was 24%.
Comparing the two c~ c the methane losses are high in both cases, beeause the process was not ~l ;. . .: . A In practise, a two-stage system should be used to reduce the methane loss and improve S the p~ e~te L~ l sulfide eoneentration. It is elear, however, that the methane loss, p~
con~.~LI pç~ ', volume and membrane area are all more f~vOl 1~1~ if the more earbon-dioxide-sc1c~ m~ l~lc is used.
FY~nple 38. Moderate amounts of earbon dioxide: lllud. .~.t., amounts of l.~Lugcl~ sulfide A one-stage lll.,n~ proeess was l~ci~ to handle a gas stream u~ -.g 100 ppm hydrogen sulfide, 10 vol% carbon dioxide and the l~ me1h~ne at a feed pressure of 6,897 kPa (1,000 psia). The target was to just meet pipeline ~ r~ of 2 vol% for earbon dioxide, wilhout col~tlu~ lg the L~u~ sulfide ~llce.ll~ - in the residue stream. The process ~ ~' . is as shown in Figure 2, where numeral 1 indieates the bark of ll.e..,l,- - myllllrc and the feed, residue and pl streams are i t' - ~ed by numerals 2, 3 and 4 l~ . The process was s~csurne~l to use one bank of 15 more carbon-dio~ide-scl~li~e ~1l~,lll1~l ~~ having the following cLal_ i~li....
Hydrogen sulrld~l~ , s.,l~livilr. 25 Carbon dioxidc/illclL ,c scl~li~ily. 20 Methane flux: 7 5 x 10~ cm3(STP)/~,.. 7-s .,.. Hg The c~ s~ nc and flow rates of the ~ and residue streams were ~ and are given in Table 22.

Flowrate (Nm3/min) 28.3 (l~OOOsefm~) 23.4 (828sefrn) 4.9 (172sefm) CH4eonc. (vol/O) 89.99 97.99 51.41 CO2eone. (vol/0) 10.0 2.0 48.54 H2Sconc. (ppm) 100 14 516 As can be seen from the table, the residue strearn, which still eontains 14 ppm, does not meet pipeline s~;fiedtion for h~lrùgel sulfide.
The proeess design c~le~ was repeated using more Il~Lu~ sulfide-scl~li~. membranes ~0 95/11738 2 1 7 ~ 3 ~ ~ PCr/US94/12099 having the following cl.~ ~tiG$-Hydrogen sulfideh.~- Ih- ~ s~l~liv;l~f: 50 Carbon dioxidc/ ' ~ selc.livil~: 13 S Mcll.anc flux: 7.5 x 10~ em3(STP)/~ s .,.. Hg The c ~ ~s~ nC and flow rates of the permeate and residue streams were ealculated and are given in Table 23.

STREAM FEED RESIDUE PERMEATE
Flowrate (Nm3/min) 28.3 (l,OOOscfm) 22.2 t783scfm) 6.1 (217scfm) CH4conc. (vol%) 89.99 97.99 61.05 C02conc. (vol%) 10.0 2.0 38.91 H2Sconc. (ppm) 100 <1 46 In this case, altho~g,h the residue stream meets the 2 vol% carbon dioxide specification, the hydrogen sulfide content ofthe streamhas beenreduced to just below 1 ppm. This O.~ results in a high n.clh_~e Ioss of 15%.
The two calculations wae repeated, using the L~J~ug~ sulfide ~-c :fi~ .. of 4 ppm as the target, but without ~I~ ulg the earbon dioxide co..c~ in the residue stream. The results for the more more carbon~ membrane are given in Table 24, and for the more l.~L og.,..-sulrlde-Sel~li~ C, UIC~ .e are given in Table 25.

STREAM FEED RESIDUE PERMEATE
Flowrate (Nm3/min) 28.3 (l,OOOscfin) 21.6 (764scfm) 6.7 (236sefm) CH4conc. (vol%) 89.99 99.28 59.90 C02conc. (vol%) 10.0 0.72 48.05 H2Sconc. (ppm) 100 4 411 As can be seen from the table, the residue stream, alfhol~ it meets the 4 ppm hydogen sulfide .~ r.~ contains only 0.7 vol% earbon dioxide. This substantial o~ ,oc~:-.g results in a high P - loss of 16%.

WO 95/11738 ~ 1 7 ~ 3 4 7 PC~rtUS94tl2099 STREAM FEED RESIDUE PERMEATE
Flowrate (Nm3/min) 28.3 (1,OOOscfm) 24.1 (8SOscfm) 4.2 (150scfm) CH~conc. (vol/O) 89.99 96.20 54.84 C02conc. (vol/O) 10.0 3.8 45.1 H2Sconc. (ppm) 100 4 643 In this ca c, the residue stream, which still contains nearly 4 vol% carbon dioxide, does not meet the pipeline cl~ :fi~ for carbon dioxide.
The ~ were repeated, using a combination proccss design as in Figure 6, where numeral 23 ;..-i c ~',5 a more L~.l,u~..-sulfide-scl~li~, bank of l,le,ubl - modules and numeral 26 - ' tt a more carbon~ le selective bank of 1l~ modules. The inr~ming gas stream 22 is at 6,897 kPa (1,000 psia). The residue stream 24 from the first bank of modules fonns the feed to the second bank.
The morc Ly~l.u~;~,l-sulfide-scl~~ bl~lC was ~Ccum~i to have thc following cl~a~t~
Hydrogen sulfidcl ' - scl~liYily; 50 Water v ~ ,lL ~C SCI&~iYity: 1,000 Carbon d;~ iCJnl~ CSCl~livi~: 13 McLlld.le flux: 7.5 x 10~ cm3(STP)/_Ill~ s ~,lllHg The more carbon~- ' -scl~li~ wlLl IC was ~cclm~rA to have the following chau~.~,t~ LL ,.
Carbon d;u~idcJ'lll.,lhdne scl~liY;Iy: 20 Hy~ Og~-lSUIrld~ l~lC scl~livity: 25 Water Y~ ' - scl~livily: 200 Methane flux: 7.5 x 10~ cm3(STP)/. Ill~-s .,l.lHg The c~ c and nOw rates of the p - -- ,t~, and rcsiduc streams from each bank of modules wcre ç~ and are given in Table 26.

~0 95/11738 2 1~ 4 3 4 7 PCTIUS94/12099 STREAM ¦ FEED ¦ RESIDUE ¦ PERMEATE
FIRST MODULE BANK (more l.~ u~ scle~ , membrane) Flowrate (Nm3/min) 28.3 (l,OOOscfm) 25.5 (9OOscfm) 2.8 (99.7scfm) S CH4conc. (vol%) 89.99 94.38 50.35 C02conc. (vol%) 10.0 5.62 49.57 H2Sconc. (ppm) 100 14 876 SECOND MODULE BANK (morc carbon-dioxide-sel~li~_ membrane) Flowrate (Nm3/min) 25.5 (9OOscfm) 23.0 (812scfm) 2.5 (88.4scfm) CH4conc. (vol%) 94.38 97.98 61.35 CO.conc. (vol%) 5.62 2.01 38.65 H2S conc. (ppm) 14 4 106 The combination process p.,- ru....C better than either of the single ~ll~..lll,l - pr~~ in this lS Co,-")o~ range. The total membrane arca used is about 135 m2. Residue stream 27 from the second stage meets pipeline 51~Y r.. ~;m~C If the ~ ...~ ~t~, streams 25 and 28 from the two banks of membrane modules are pooled, the permeate c~n~ is 510 ppm h~ug_., sulfide, 44 vol% carbon dioxide and 56 vol% mtolh~ne The methane loss in the pooled ~ ...- ,t~ C iS about 11.5%. This loss could be reduced if the process were o~
20 ~
r u , ~c s 39 and 40 sho~v ll~dtn~ll lJIU~ in which the membrane process does not bring the gas stream to pipeline s, f - for all co--~
FYa~nrle 39. Mc.,.l,al,c plus s~,a~ 6;.lA6 proccss A process was designed tohandle a gas stream c~ -- -.g 1,000 ppm h~og~.. sulfide, 0.1 vol%
water vapor, 4 vol% carbon dioxide and the r~n~intler methane, at a feed proesure of 6,897 kPa (1,000 psia). The process includes a one-stage membrane separation step, followed by a sca~ 5..1g step to bring the hyd~u~l sulfide o~.lll down further to 4 ppm. The Sca~ g step could be carried out using an iron sponge, for example. The process was ~cs~med to use one bank of more hy~u~;~,.l-sulfidc-sclc~ membranes having the following e' t~
30 Hydrogen sulfidcJl ' - sel~livity. 80 Watervapol/ ..~ c sel~livily. 1,000 Carbon dioxid~ & sel~livily. 12 M~ flux: 1 x 10~ cm3(STP)/~,.. l'-s .,.. IHg wo95/11738 2~ ~3 ~ so Pcrluss4/l2099 The c~ nc and flow rates of the permeate and residue streams were r~lr~ t~ and are given in Table 27.

S STREAM FEED RESIDUE PERMEATE
Flowrate (Nm31min) 2.8 (lOOscfm) 2.S (90.3scfm) 0.3 (9.7scfm) CH4conc. (vol/O) 95.8 98 74.9 C02conc. (vol/O) 4.0 1.9 23.1 H2Sconc. (ppm) 1,000 40 990 Watervaporconc.(vol/O) 0.1 2ppm 1.0 The.~ areausedwas~ tobeabout70m2. Thestagecutwas justunder 10%and the methane loss into the ,~ was 7.6%. The process l~ùdu~s a residue stream that meets pipeline Y; r.~ for carbon dioxide and water vapor, but needs furlher polishing to remove h, ~ U~ll sulfide.
lS
Example 40. Process ;,~ amine plant for h~dluEj~,,. sulfide removal A process was ~lesi~ed to handle a gas stream c~ -: .g O.S vol % LyJIug.,,~ sulfide, 20 vol%
carbon dioxide and the ,, ' m~th -~, at a feed pressure of 6,897 kPa (1,000 psia). The process uses a one-stage Ill.,.llbl - s-r step to car~y out a first a~lJ~ ' of carbon dioxide and L~l~ug~
20 sulfide, followed by an amine plant to bring the stream to pipeline a~ '~ The process was acc~ oA
to use one bank of more Lydlug_.l-sulfidc-sel~~ having the following cL __h.i~ ,a.
IIy~Luge,l sulfidc/l ' - s.,l~liv;ly. SO
CarbonLo.~idc/~ ' - sdc.liv;ly. 13 2S Methane flux: 7.S x 10~ cm3(STP)/~.. l7-s-.,.. Hg The co...~ ;o..c and flow rates of the ~ ~ and residue streams wGre ~-' ' d and are given in Table 28.

STREAM FEED RESIDUE PERMEATE
Flowrate (Nm31min) 28.3 (l,OOOscfm) 23.8 (840scfin) 4.5 (160scfin) CH4conc. (vol/O) 795 88.8 30 C02conc. (vol/O) 20 11.1 67 H2Sconc. (vol/O) 0.5 0.05 2.9 ~JO 95/11738 ~ ~ 7 4 3 4 ~ PCT/US94/12099 Sl The l"w-L,~c area used was ~ ~s ' to be about 70 m2. The stage cut was just under 16% and the methane loss into the permeate was 6%. The process p~ u~s a residue stream from which 90% of the h~llU~W~ sullide and about 50% of the carbon dioxide has been r~ u.~. This residue stream passes to the amine plant for additional l-~t~ to bring it within specification for carbon dioxide and hydrogen S sulfide.

Claims (14)

We claim:

1. A membrane process for treating a gas stream comprising hydrogen sulfide, carbon dioxide and methane, said process comprising the following steps:

(a) providing a feed stream containing carbon dioxide in a concentration less than about 3% to less than about 10% and hydrogen sulfide in a concentration more than about 10 ppm to more than about 300 ppm, with the lower end of the carbon dioxide range corresponding to the lower end of the hydrogen sulfide range (<3% carbon dioxide; >10 ppm hydrogen sulfide) and the upper end of the carbon dioxide range corresponding to the upper end of the hydrogen sulfide range (<10% carbon dioxide; >300 ppm hydrogen sulfide), (b) passing said feed stream through a membrane unit containing a membrane characterized by a selectivity for hydrogen sulfide over methane of at least 35 and a selectivity for carbon dioxide over methane of at least 12, said selectivity being measured with a mixed gas stream containing at least hydrogen sulfide, carbon dioxide and methane and at a feed pressure of at least 500 psig;

(c) withdrawing from said membrane unit a residue stream containing carbon dioxide in a concentration no greater than about 3 vol% and hydrogen sulfide in a concentration no greater than about 20 ppm.

2. A membrane process for treating a gas stream comprising hydrogen sulfide, carbon dioxide and methane, said process comprising the following steps:

(a) providing a feed stream containing carbon dioxide in a concentration less than about 10% to less than about 20% and hydrogen sulfide in a concentration more than about 300 ppm to more than about 600 ppm, with the lower end of the carbon dioxide range corresponding to the lower end of the hydrogen sulfide range (<10% carbon dioxide; >300 ppm hydrogen sulfide) and the upper end of the carbon dioxide range corresponding to the upper end of the hydrogen sulfide range (<20% carbon dioxide; >600 ppm hydrogen sulfide);

(b) passing said feed stream through a membrane unit containing a membrane characterized by a selectivity for hydrogen sulfide over methane of at least 35 and a selectivity for carbon dioxide over methane of at least 12, said selectivity being measured with a mixed gas stream containing at least hydrogen sulfide, carbon dioxide and methane and at a feed pressure of at least 500 psig;
(c) withdrawing from said membrane unit a residue stream containing carbon dioxide in a concentration no greater than about 3 vol% and hydrogen sulfide in a concentration no greater than about 20 ppm.

3. A membrane process for treating a gas stream comprising hydrogen sulfide, carbon dioxide and methane, said process comprising the following steps:

(a) providing a feed stream containing carbon dioxide in a concentration less than about 20% to less than about 40°/0 and hydrogen sulfide in a concentration more than about 600 ppm to more than about 1%, with the lower end of the carbon dioxide range corresponding to the lower end of the hydrogen sulfide range (<20% carbon dioxide; >600 ppm hydrogen sulfide) and the upper end of the carbon dioxide range corresponding to the upper end of the hydrogen sulfide range (<40% carbon dioxide; >1% hydrogen sulfide);

(b) passing said feed stream through a membrane unit containing a membrane characterized by a selectivity for hydrogen sulfide over methane of at least 35 and a selectivity for carbon dioxide over methane of at least 12, said selectivity being measured with a mixed gas stream containing at least hydrogen sulfide, carbon dioxide and methane and at a feed pressure of at least 500 psig;

(c) withdrawing from said membrane unit a residue stream containing carbon dioxide in a concentration no greater than about 3 vol% and hydrogen sulfide in a concentration no greater than about 20 ppm.

4. The process of claim 1, 2, or 3, wherein said selectivity for hydrogen sulfide over methane is at least 50.

5. The process of claim 1, 2, or 3, wherein said feed pressure at which said selectivity can be obtained is at least 1,000 psig.

6. The process of claim 1, 2, or 3, wherein said residue stream contains carbon dioxide in a concentration no greater than about 2 vol%.

7. The process of claim 1, 2, or 3, wherein said residue stream contains hydrogen sulfide in a concentration no greater than about 4 ppm.

8. The process of claim 1, 2, or 3, wherein said membrane comprises a composite membrane having a selective layer comprising a polymer that is rubbery under the operating conditions of the process.

9. The process of claim 1, 2, or 3, wherein said membrane comprises a block copolymer containing a polyether block.

10. The process of claim 1, 2, or 3, wherein said membrane comprises a polyamide-polyether block copolymer having the general formula wherein PA is a polyamide group, PE is a polyether group and n is a positive integer.

11. The process of claim 1, 2, or 3, wherein said feed stream comprises natural gas.

12. The process of claim 1, 2, or 3, further comprising:

(d) withdrawing from said membrane unit a permeate stream enriched in carbon dioxide and hydrogen sulfide and having a methane content such that methane loss from said feed stream is no more than about 5%.

13. The process of claim 12, wherein said methane loss is no more than about 2%.

14. The process of claim 1, 2, or 3, wherein said feed stream contains carbon dioxide, dioxide sulfide and water vapor, all in concentrations above pipeline specification, and wherein said residue stream meets pipeline specifications for carbon dioxide, hydrogen sulfide and water vapor.