CA2148609A1 - Gas separation process - Google Patents

Gas separation process

Info

Publication number
CA2148609A1
CA2148609A1 CA002148609A CA2148609A CA2148609A1 CA 2148609 A1 CA2148609 A1 CA 2148609A1 CA 002148609 A CA002148609 A CA 002148609A CA 2148609 A CA2148609 A CA 2148609A CA 2148609 A1 CA2148609 A1 CA 2148609A1
Authority
CA
Canada
Prior art keywords
gas
membrane
selectivity
methane
stream
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Abandoned
Application number
CA002148609A
Other languages
French (fr)
Inventor
Lora G. Toy
Ingo Pinnau
Richard W. Baker
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Membrane Technology and Research Inc
Original Assignee
Individual
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Priority claimed from US07/971,331 external-priority patent/US5281255A/en
Application filed by Individual filed Critical Individual
Publication of CA2148609A1 publication Critical patent/CA2148609A1/en
Abandoned legal-status Critical Current

Links

Classifications

    • BPERFORMING OPERATIONS; TRANSPORTING
    • B01PHYSICAL OR CHEMICAL PROCESSES OR APPARATUS IN GENERAL
    • B01DSEPARATION
    • B01D53/00Separation of gases or vapours; Recovering vapours of volatile solvents from gases; Chemical or biological purification of waste gases, e.g. engine exhaust gases, smoke, fumes, flue gases, aerosols
    • B01D53/22Separation of gases or vapours; Recovering vapours of volatile solvents from gases; Chemical or biological purification of waste gases, e.g. engine exhaust gases, smoke, fumes, flue gases, aerosols by diffusion
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B01PHYSICAL OR CHEMICAL PROCESSES OR APPARATUS IN GENERAL
    • B01DSEPARATION
    • B01D69/00Semi-permeable membranes for separation processes or apparatus characterised by their form, structure or properties; Manufacturing processes specially adapted therefor
    • B01D69/02Semi-permeable membranes for separation processes or apparatus characterised by their form, structure or properties; Manufacturing processes specially adapted therefor characterised by their properties
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B01PHYSICAL OR CHEMICAL PROCESSES OR APPARATUS IN GENERAL
    • B01DSEPARATION
    • B01D71/00Semi-permeable membranes for separation processes or apparatus characterised by the material; Manufacturing processes specially adapted therefor
    • B01D71/06Organic material
    • B01D71/44Polymers obtained by reactions only involving carbon-to-carbon unsaturated bonds, not provided for in a single one of groups B01D71/26-B01D71/42

Landscapes

  • Chemical & Material Sciences (AREA)
  • Chemical Kinetics & Catalysis (AREA)
  • Engineering & Computer Science (AREA)
  • Analytical Chemistry (AREA)
  • General Chemical & Material Sciences (AREA)
  • Oil, Petroleum & Natural Gas (AREA)
  • Separation Using Semi-Permeable Membranes (AREA)

Abstract

A process for separating condensable organic components from gas streams (1). The process makes use of a membrane (2) made from a polymer material that is glassy and that has an unusually high free volume within the polymer material. Condensable organic components are removed as permeate stream (4), while a residue non-permeate stream is removed as stream (3).

Description

,.,1 WO 94/098~6 i~ 1 4 8 ~j ~ 9 P~/~Sg3~10~3~

~AS-SEPARATION PROCESS
This invention was made with Government support under Contract Number DE-FG03-9OER810fi6, awarded by the Department of EDergy. The Go~ernment has certainrights în this invention.
S FIEI.D l:)F THE INVENTION
The invention relates ~o a gas-separation process. More particularly, the invention concerns the remoYal of condensable organic vapors from gas mixt~res, especially the removal of hydrocarbons from gas mixtures.
BACKGROUND OF THE INVENTION
Gas-separation membranes are known and are in use in such areas as production -;, of oxygen-enriched air, production of nitrogen for blanketing and other applications, separation of carbon dioxide from methane, hydrogen recovery from Yaribus gas mixtures and removal of organic vapors from air.
The optimum separation membrane for use in any gas-separation application com~ines high selectivity with high flux. Thus the membrane-makin~ industry has ~1 ~ engaged in an ongoing quest for membranes with improved flux/selectiYity performance.
Gas and vapor permeation through polymer membranes is usually rationalized by the solution-diffusion model. This model assumes that the gas phases on either side of the membrane are in thermodynamic equilibrium with their respe~tive poJymer interfaces, and that the interfacial sorption and desorption process is sapid compared with the rate of diffusion through the membrane. Thus ~he rate-limitin~ step is diffusion through the ~ polymer membrane, which is governed by Fiek's law of ditfusion. In simple cases, Fick's ¦ ~ law leads to ehe equation J/Qp) D (D S)/~ , ( ) where Q is the pressure-normalized flux lcm~(STP)/cm2 s cmHgl, J is the volumetric flux ~ ~ 25 per membrano area [cm~(STP)/cm2 sl, D IS the diffusion eaefficient of the ~as or vapor in j ~ the membrane [cm2/s~ and is a me~sure of the ~as mobility, e is the membrane thickness, is the Henry's l~w sorption coefficient linking the concentration of the gas or vapor in ~he membrane material to the pressure in the adjaeent gas [cm~(STP~/cms cmHg~, and Ap is the pressure difference across the membrarle. The product 17 S can aJso be expressed as ::
:

WO 94/09886 ~ 3 9 Pcr~usg3/lo736 the permeability coefficient, P, a measure of the ra~e at which a particular gas or vapor moves through a membrane of standard thickness (I cm) under a standard pressure differ~nce (I cmHg). As can be seen from Equation 1, the pressure-normalized flux is inversely proportional to the membrane thickness.
S For a given membrane material, the ideal selectivity, ~AB~ for gas A over B is defined as the ratio of the permeability coefficien~s of the gases:

" PAIPJ ' t~hlD~ (SklS"~ , ~2) where PA and PB are the permeability coefficients of gases A and B, respectively, as determined from the measured pressure-normalized fluxes of two gases, the fluxes being measured separately, each with a pure gas sample, through a defect-free membrane sample ~', 10 of the same thickness, and being expressed in cm~(STP)/cm2 s cmH~ or other consistent ~ 1 ~1 units. Selec~ivity, as defined in Equation 2, is a product of two terms. The first term is 3 the ratio of the diffusion coefficients and is usually called the mobility selectivity. This I term reflects ehe relative size of the permeants. ln the case of the separation of organic ; compounds from permanent gases, such as nitrogen, the diffusion coefficient of the organic vapo~ is always less than that of ni~rogen, so the mobility selectivity term is less than one. The second term is the sorp~ion selectivity and reflects ~he relative sorption of the two permeants. In general, the more condensable the component, the higher its sorp~ion. Thus, in the case of the separation of a more condensable organic compound ¦ ~ from a permanent gas or less condensable organic or inorganic compound, the sorption 1 20 selectiYity term will usually be greater than one. Whether a particular membrane material is selec~ive for the rnore c~ndensable components of a gas mixture depends on the balaDce of these two terms for that mate~ial.
Diffusion coefficients are generally several orders of magnitude higher in rubbery .
polymers than in glassy polymers and are substant;ally less dependent on the penetrant size, particularly ;n the case of large, condensable molecules. As a result, the selectivity of .
rubbery polymers is mainly determined by the sorption term and ru~bery materials are , usually condensable selectiYe. G}assy polymer selectivities, on the other hand, are dominated by the diffus~on term and glassy polymers are usually gas-s~lective. Data illustrating the standard 'oehavior of rubbery and glassy polymers are shown in Figure 2, originally prepared by the German company, (;KSS. Only rubbery polymers, therefore, .j ~:
WO 94/098~S 2 1 ~ ~ 6 ~ 9 PC~/US93/10736 ,~ ) have been considered useful for separating condensable organic compounds from other gases and vapors.
In recent years, some polymer materials with unusua]ly high permeabilities have been synthesized. Perhaps the bes~ ~.nown of these, and representative of the c}ass, is S pclytrimethylsilylpropyne (PTMSP), a polymer synthesized by T. Masuda et al. in Japan.
Although PTMSP is glassy, up to at least about 200C, it exhibits an o~cygen permeability of 10,000 Barrer or above, more than 15 times higher than that of silicone rubber, previously the most permeable polymer Icnown. The selecti~ity for o~ygen/nitrogen, however, is low (15-1.8) The high permoability appeàrs to be associated with an unusually high free-volume within the polymer material, and has been confirmed with many examples of pure gases and ~apors, including oxygen, nitrogen, hydrogen, helium, methane, ethane, propane, butane and higher hydrocarbons, sulfur hexafluoride and carbon dioxide. These pur~-gas data suggest that PTMSP will exhibit poor selectivity for most gas separations. For example, a paper by N.A. Plate et al. (nGas and vapor ; ¦ 1~ permeation and sorption in poly(trimethylsilylpropynen, Journal of Membrane Science, Vol.
j ~ 6û, pages 13-24, 1991) gives polymer permeabilities of 2.6 x 10-7cmS(STP).cmjcm2 s cmHg~ for oxygen~ and 1.~ x~ 10 ~7 cm~tSTP).cm/em2 s cmHg for nitrogen, giving a ~alculated selectivity of 1.7. The same reference gives polymer permeabilities of 2.7 x 10-7 cm~STP).cm/cm2 s cmHg îor methane, 1.9 x 10-7 2Q cm~(STP).cm/cm2 s cmlHg for propane and 2.3 x 10-7 cm~STP).cm/cm2 s cmHg~ for n-~ .
butane, giving a c~lculated selectiYity for propane/methsne of 0.7 and for butane/methane of 0.85. A paper by M. Langsam et al. (nSubstituted Propyne Polymers. I. Chemical surf~ce ~ modification of poly~l-(trimethylsilyl)propyne] for gas sepRration membranesn, Qas ;~ ~ Separation~and Purification, Vol. 2, pages 162-170, 1988) gives a carbon dioxide/methane i I ~selectiYity for PTMSP of ;2.07, compared witk 9.56 fo~ silicone rubber. A pa~er by K.
Takada et al. (nGas Permeab;lity of Polyacetylene~ Carryin~ Substituents". Journal of Applied Polym~er Science9 Vol. 30, pages 1605-1616, 1985) includes the statement that:
"Very interestingly, poly~l-(trimsthylsilyl)-l-propyne] ~films show permeabilitycoefficients as~high as 10-7-10-8 ~o every gas. However, permselectivities of these films ~; 30 ~ for two diff~rent gases aro relatively poor." Thus the m~terial was characterized, at least initially, as~ of great~ academic interest, because of its e~traordinary permeability, but exhibiting selectivities too low for commercial use. ~
: : :
:
1 ~

WO 94/09886 Z 1 a~ 8 ~ ~ ~ P~/US93~10736 As soon as its remar~;able permeability properties were announced, PTMSP
at~racted attention from the membrane community at large. A nurnber of experimenters reported that the permeation properties of PTMSP appear to be unstable over time, raising further douhts as to the usefulness of the materi~l for practical applications. In particular, the oxygen permeability was found to drop dramatically. For example, Masuda et al. found that the oxygen permeabili~y fell to about 1% of its orginal value when the membrane was left at room temperature for several months.
More recently, the consensus of opinion in the art has been that the loss in permeability arises primarily from sorption of volatile materials from the environment of the rnembrane. If the membrane is mounted in a system containing a vacuum pU~Dp, for example, vaporized or aerosol vacuum oil may be sorbed into the membrane material. A
similar effect may occur if the membrane is simply standing in the air for prolonged periods. For example, a paper by T. Nakagawa et al. ("Polyacetylene deriYatives as membranes for gas separation", Gas Separation and Purification, Vol. 2, pages 3-8, 1988) states that "the PMSP membrane showe~ strong affinity to volatile materials. It was considered that, in addition to the thermal hysteresis, the reason for unstable gas permeability is the ad~orption of volatile materials existing in the atmosphere.~
This property has been turned to advantage by several workers. For ex~mple, the above-cited Nakagawa paper also discusses the performance of PTMSP membranes ehat haYe been deliberately exposed to a variety of additives, including dioctyl phthlate (DOP) and polyethylene glycol (PEG). The treated membranes exhibited permea~ion properties stable over time, and, although the oxygen permeability was reduced from 8,000 Barrer to about 300-400 Barrer, the o%~gen/~itrogen selectivity improved from 1.6 to 3.3, renclering the membranes "prospectiYe as membranes for oxygen enrichment". Sirnilar 25 I results have been reported by S. Asakawa et al. ("Composite membrane of polyll-(trimethylsilyl)-propyne~ as a potential oxygen separation snembrane", Gas Separ~tion and Purification, Vol. 3, pages 117-122, 1989), who apparently produced stable PTMSPmembranes by coating the PTMSP layer with a protective layer of silicone rubber9 and who also concluded that, "This membrane, therefore, may be promising for industrial oxygen separation." M. Langsam et 21. (U.S. Pa~ent 4,859~21~, August 22, 19~9, assigned to Air Products and Chemicals, Inc.) added Nujol oil, silicone oil or ethylene oxide-based surfactants to the casting solution when preparing PTMSP membranes. The membranes WO 94/09886 214 ~ ~ 3 ~ PCI/USg3/10736 ., , ' ' !3, showed permeation properties stable o~er time, reduced permeabilities and improved selectivities for oxygen/nitrogen and carbon dioxide/nitrogen.
Other attempts to modify the materia1 to increase its selectivity have been made.
For example, U.S. Patent 4,657,~64, to M. Langsam, assigned to Air Products and S Chemicals, Inc., describes a surface fluorination ~echnique that increases theoxygen/nitrogen selectivity by at least ~0~ o~er its unmodified value. Thus, use of the material has focued on o~cygen/nitrogen separation, and ways in which the ex~raordinary oxygen permeability can be preserved yet the low o~yg0n/nitrogen selectîvity enhanced.

Almost all of the permeation data that have been published concern pure gas ;i .
10 experiments. However, a study by S.R. AUY;I et al. ("Mechanisms of gas transport in poly(l-trimethylsilyl-l propyne),PolymerPreprints,Vol.32(3),pages380-383, 1991)was carried out using mixtures of a heavy gas ~carbon dioxide or sulfur hexafluoride) and a light gas (helium or nitrogen). Th~ study showed that the permeability of the light gas is substaneially reduced in the presence of the heavy g~s. It was postulated that the heavy gas 15 is adsorbed onto surfaces of voids within the structur~ of the polymer and may be transported through the material by surface diffusion, and further that these surface layers may build up and block diffusion of the light gas through the void areas. The net result was an increase in the solectivity for the heavy gas over the light gas when measured with gas mi~ttures rather than calculated from pure gas permeabiiities.
To summarize the above discussion, it may be seen that glassy, high-frse-volume polymers, of wh;ch PTMSP is the most widely studied example, exhibit unusual gastraniport properties. These properties do not conform ~to, and do not appear to follow from the standard solution/diffusion model of gas transpo~rt.~ Furth~rmore, the properties are affected in a not fully understood ~fashion by sorption of a Yariety of v~latile materials.
Behavior with mixed Sases has not been s~udied, except in a very limited way, but the results obtained again are inconsistent wit~ those obtained from cQnvetltional polymer ~; ~ materials. ~ ~ ~
Natural gas is very important both as fuel and as a basie industrial raw material.
The composition of raw na~ural gas varies from field to field. It may Gontain more than 95% methane with small amounts of other hydrocarbons, nitrogen, carbon dioxide, hydrogen sulfide, or water vapor. On the other hand, it may contain up to 15% ethane, WO 94/09~86 21~ ~ G ~ PCI/~S93/10736 ~' 6 propane, butane, or combinations of these components. For safety reasons, the Btu rating of natural gas to be carried through a pipeline is controlled within a narrow range (950 -1,050 Btu/ftS). However9 because of the higher Btu values of ethane, propane, butane, and pentane, natural ~as streams containing significant amounts of these components are too high in Btu value to be fed directly to a pipeline or to be used directly as fuel. Equally important, higher hydrocarbons are too valuable industrially to be burned as secondary ;~ components in gas. Thus, the natural gas stream must undergo some form of treatment to remove undesirable components, to ~ring the Btu value to the s~andard level, and to recover the valuable hi~her hydrocarbons. U.S. Patent A,857,078 discloses a process for removing propane, butane and higher hydrocarbons from natural ~as using a composite membrane with an ultrathin rubbery selective layer.
SUMMARY OF THE INVENTION
The inYention is a process for separating a condensable organ.ic component from a gas stream. The process involves running the gas stream containing thP condensable organic component across a membrane that is selectively permeable to that csmponent.
The condensable component is therefore concentrated in the stream permeating the: ' membrane; the residue, non-permeating, stream is correspondingly depleted in condensable content. The process differs from processes previously u~ed for separating condensable organic components from gas streams in the nature of the membrane that is used.
The membrane is made from a polymer material characterized as follows:
l. Glassy 2. Unusually high free volume within the polymer material . 3. Pure gas data:suggest ~oor selectivity : ~ 4. Measured mixed gas selectivity is substantially better than calculated pure gas selectivity 5. Selectivity of material depends on thickness The membrane material has characteristics and exhibits prsperties that are ¦: ~ fundamentally different from those of the me,~branes previously used for this type of i separa~on.
~: ~ 30 Because the materials are glassy and rigid, an unsupported film of the polymer may j ~ often be usable as a single-layer gas separation membrane. Alternatively, the separation 'l membrane may be a layer that forms part of a thicker structure, such as an asymmetric '~2 ! wo ~4/098g6 2 i 4 8 ~ ~ ~ PCl~/US93/1073~

membrane or a composi~e membrane.
The driving force for permeation across the membrane is the pressure difference between the feed and permeate sides, which can be generated in a variety of ways. The membrane separation process produces a permeate stream enriched in the condensabte component compared with the feed and a residu~ stream depleted in ~he condensable component.
The membrane separation process may be configured in many possible ways, and may include a single membrane unit or an array of two or more units in series or cascade arrangements. Eighty to 90% or above removal of the condensable content of the feed to the membrane system can ~ypically be achieved with an appropriately designed membrane separation process.
The process is useful in separating any condensable organic compound from air, permanent gases, or less condensable organic or inorganic compounds. The process has advantages over processes that were previously used for such separations in that it combines high flux of ~he condensable component with unexpectedly high selectivity for ~he condensable component.
.
In one specific exemplary embodiment, the process of the in-vention can be used for separating C~+ hydrocarbons from natural gas. Natura1 gas typically has methane as its major component, and may also con~ain signiîicant ~uantities of ethanet propane, butane 20~ and other hipher hydrocarbons, nitrogen, carbon dioxide, water ~rapor, and hydro~en sulfide. In this case, the process inYolves running the na~ural gas s~ream across a membrane thnt is selectively permeable to C~ hydrocarbons over methane. The higher hydrocarbons are therefore concentrated in the stream permeating the membrane; the residue, non-permeating9 stream is correspondingly depleted in hi~her hydrocarbons. The 25 I ~process differs from processes previously used for separating C~ hydrocarbons from natural gas in the n ature of the membrane shat is used.
is to be understood ~hat the above summary and the following detailed description are intended to explain and illustra~e the invention wi~hout restricting its scope.
BRIEF DESCRIPTION OF THE DRAWINGS
Figure I is a schematic drawing of the process of the invention in its simplest form.
Figure 2 is a graph of permeability as a function of molecular si~e, expressed as critical YOIUme~ for a typical rubbery and a typical gla~sy polymer.

.
`
:
:: :

i;~

;~ WO 94/09886 PCr/US93/10736 2148~ i i 8 Figure 3 is a graph of pure gas butane and methane pressure-normalize f luxes plotted as ~ a function of inverse separation membrane thickness.
`:~
Figure 4 is a graph of ideal butane/methane selectivity as a function of separation ,~ membrane thickness.
I ~ Figure ~ is a graph o~ pressure-normalized flu~es plotted as a function of inverse separatioD membrane thickness. The fluxes were measured with a gas mixture consisting ! of 86% methane, 1û% ethane, 3% propane and 1% butane.
Figure 6 is a graph of mixed gas butane/methane selectivity plotted as a function of ,1 in~erse separation membrane thiclcness. The selecti~rity w as measured with a gas mix~ure ;~ 10 consisting of 86% methane, 10% ethane9 3% propane and 1% butane.
DETAILED DESCRIPTION OF THE IN~ENTION
The term membrane unit as used herein means one or more membrane modules arranged in parallel, so that a portion of the incoming gas stream passes through each one.
The term series arrangement means an arrangemellt of membrane modules or units ~5 connected together such that the residue stream from one module or unit becomes ~he feedstream for the next. The term cascade arrangement means an arran~ement of membrane modules or units connected together such that the permeatç stream from one module or unit becomes the feedstream for the next.
The term C~+ hydrocarbon(s) is an abbreviation meaning any hydrocarbon havir.g `~ 20 three or more carbun atoms.
,, :
The process of the inYention involves running a gas stream containing at least two components, at least one of which is a condensable organic compour~d, across a membrane that is selectively~permeable to the condensable componen~ over the second component.
` The term condensable a5 used herein refers to fluids below ~beir csiticaJ temperatures, 25 I having boiling~ points greator than -~0C at atmospheric pressure. If a mixture containing two or more condensable components is to be trea~ed, ~hs term condellsable refers to the more readily condensabie component ~or comporlents. In the specific case o~ natural ~as treatment, the process of the invention învolves running a natural gas stream across a membrane that is select~vely permeable to C~+ hydrocarbons over methane, ethane and nl:rngen. ~ ~
The gas streams that may be treated by the process of the invention are diverse.By way of non-limi~ing example, many industr;21 processes produce gas streams containing ~x :~s .'r'~
~ - ~ W0 94/09886 PCI/US93/10736 ~ 21~0~ 9 organic vapors in air or nitrogen. Such organic vapors may be aliphatic or aromatic hydrocarbons, for example, or halogenated hydrocarbons, such as fully or partially substituted chlorinated hydrocarbons, fluorinated bydrocarbons and chlorofluorocarbons (CFCs and HCFCs). Streams of organic compounds in other gæes are also found. Forexample, hydrogenation reactions in the chemical industry yield off-gas streams containing hydrogen and various hydrocarbons. Mixed organic compound streams occur, particularly in chemical processing, petrochemical refining and natural gas treatment.
The process differs from processes previously used for separating condensable components from gas streams in the nature of the membrane that is used. As discussed in I Q the background section above, the conventional belief of the art is that rubbery membranes should be used when a condensable organic compound is to be separated from a gasmixture. For example, US. Patent 4,857,078, which concerns removal of C~
hydrocarbons from natural gas, states that "Glassy polymers ... are, however, relatively unselective for one hydrocarbon over another, and are unsuitable for separating me~hane or ethane from Cs or C ~*~ hydrocarbons. In fact, these types of membrane often are more permeable to methane than to the C2+ hydrocarbons.~ (column 2, lines 46-52). The patent goes on to deseribe suitable membranes for this partlcular condensable/non-condensable separation, as follows:~"The permselective membranes used in the invent;on tben ~re rubbery non-crystalline polymers, that is they have a glass transition temperature at least 20C below the normal operating temperature of the system. Thermoplastic elastomers are also useful." (column, 7, lines 8-I2~. Likewise, U.S. Patent 5,089,Q33, which covers a hybrid process including condensatioD and membrane separation for removing condensable components in gen~ral from gas streams, states that: `nTo remove sn organic vapor as the preferentially permeating component, ~a~ number of rubbery polymers could be used.
25 ~ I Examples, include n itrile rubber, ~ nieoprçne,~ silicones rubbers,j inrluding ;, polydimethylsiioxane, chlorosulfonated polyethylene, polysilicone-carbonate copolymers, ; fluoroelastomers, ~plasticized polyvinylchlorid~, ~polyurethane, cis~po}ybutadiene, cis-polyisoprç~ne, ~ ~poly~butene-l)j ~ polystyrene-butadiene copolymers, styrene/butadiene/styrene block copolymers~; and styrene/ethylene/butylene blockcopolymers. Particulariy preferred rubbers are~silicolle rubbers." (~ol. 9,1ines ~9-41).
In complete coDtrast to these teachings, the membranes that are used to separate- condensable organic components from gas streams according to the present invention are ~`
WO94/098~6 PCr~U

~ ~148~3g ' charac~erized as follows:
1. Glassy 2. Unusually high free volume within the polymer ma~erial ..;
3. Purie gas data suggest poor selectivity S 4. Measured mixed gas selectivity is subssantially better than calcuiated pure gas ~1 selectivity S. Selectivity of material depends on thickness 1. ,C~lassY
The materials that havc been found so far tQ be useful in carrying out the process of the invention have glass transition temperatures Tg at least aboYe SQC, and typically ~; much higher glass transition temperatures, such as above 100C, 200C or even higher.
~i~ Thus, they are always complete}y glassy and rigid under the conditions in whieh they are used in .he process of the invention.
2. Unusuall~hi~h free volume w thin the ~olvmer~
The materials are ~Iso characterized by unusually high free volumes, as estimated from vapor solubility data as in W.J. Koros et al., J. Membrane Science, ~ol. 2, page 165, ,;~
197~. Conventional glassy polymers typical}y: have free volumes, ~F~ within the polymer : itself of a fiew percent, such as 3% for polycarbonate, or 6% for polyphenylene oxide. The :~ materials that exhibit selectivity/thickness dependence have higher free volumes, ,, preferably more than 10q~ and most preferably more than 20%. For example, PTMSP, a silieon-containing substituted polyacetylene, has a free volume of about 2S4~ aceording to this method.
3/4.
7 ~
i In the gas-separation-membrane art, two types of selectivity are commonly reported. One is the ideal selectivity, which is the calculated ratio of the measured pressure-normalized fluxes sf two gases, the fluxes being measured separately, each with a pure ~as sample, through a defect-free membrane sample of the same thickness~ and being expressed in cm~(STP~/cm2 s-cm~Ig or other consistent units. The other i~ ~he actual or mixed-gas selectivity, measured with a gas mixl:ure containing two or more gases to be sep~rated.

~ 2:~8~3~
-~ WiO 94/098~6 PCI`~US93/10736 .,`;j~l , As discussed in the background section above, permeability data from the published li~erature indicate that the glassy, high-free-volume polymer materials usable in the process of the invention have poor ideal selectivities for more condensable organic .~ compounds over less condensable organic compounds or inorganic compounds. In fact, the data show ehat these materials may, in some cases, be slightly selective for the less s conde~sable over the more condensable component, for example, methane over propane 1'!,i, or butane.
~`~ As reported in the Examples section below9 we also found that permea~ion tests !~
performe~ with pure gas samples yielded low calculated ideal selectivities for more condensable organic compounds ov~r less condens~ble organic compounds or inorganic ;.,,~
compounds.
We found, however, that the measured mixed-gas se1ectivity for more condensable ''.~ organic compounds over less condensable organic compounds or inorganic compounds was dramatically better than ~he ideal selectivity. Typically, the mlxed gas selectivity was at least three times the ideal selectivity, and frequently more, such as five times, seven times or ten times the ideal selectivity.
.~ It is certainly not unlcnown for a gas-separatiol membrane under experimental conditions or use to exhibit a lower selectivity than the ideal selectivity, for ae least three possible reasons: :
~i) The separation membrane has one cr more defects that permit indiscriminate bulk flow of both gases, thereby lowering the selectivity. The thinner the membraQe, the more likely is this to be a problem.
(ii) The separation membrane is supported on a subs~rate that offers a resistance to gas permeation that is not ~nsignificant compared with the resistance of ~he separaition 25 I i membrane and, therefore, has an influence on the oYerall selectivity of the ¢omposite.
~ This phenomenon is discussed in detail in co-owned U.S. Patent 4,931,181.
q ~ ~ giii) The mixed-gas selectivity is inherently lower than the ideal selectivity calculated from purc gas measurements. This is a very common phenomenon, often caused by plasticization or swelling of the membrane by one component in the mixture. Condensable organic compounds are very likely to cause such an e~ect.
=
It is, however, very surprising for the mixed gas selectivjty to be better ~han the ideal selectivit~, especially in the case of the gas streams to which the process of the WO ~4/~9~i86 PCI/US9~/10736 invention can be applied.
. SelectivitY of material de~ends on thickness The materials that have been found so far to be useful in carrying out the process of the invention possess a sslrprising and hitherto totally unsl~spected property, in that, `, ~i' 5 when an isotropic, essentially defect-free film of the material is used to separate one gas from another, the separation selectivity exhibited by the film depends on its thickness.
~j The actual selectivity increases, up to a maximum value, as film thickness increases. To '~.
~!~ applicant's knowledge, this unexpected behavior has never been observed previously with any other polymer materials. The reason why this behavior is observed is not yet known, but it appears that gas transport is not conforming completely to the solution/diffusion model that is the standard approach for understanding gas transport through dense polymer films. This attribute can be used in tailoring membrane performance to suit specific applications.
¦ Non~limiting examples of the types of polymer material that fall within the 1 1 j definition of glassy, high-free-volume materials useful for carrying out the process of the invention include:
(i) Substituted acetylenes, ha~!ing the general structural formula Rt { C C }m . I

where R~ and R2 are independently hydrogen, a halogen, C~E; or a linear or branched Cl-C4 alkyl group, (ii) Silicon-containing polyace~ylenes, ha~ing the general structural formula R l ~{-- C = C--}
m .

~n `R2_ Si _ R4 where R1 is a linear or branched C~-C4 alkyl group, R2 and R~ are independently linear :~j ) WC~ 94/09886 2 1 ~ ~ 5 ~ 9 PCI/US93/10736 ~ or branched Cl-C6 alkyl groups, R4 iS a linear or branched Cl-C12 alkyl or aryl group, and j~ X is a Cl-C~ alkyl group.
Germanium-containing polyacetylenes, having the general structural formula R
'i` . . I
{--- C = C --}

' ~X)n ;j R2_ Ge - R4 `

~` . where R1 is a linear or branched C1-C4 alkyl group, R2 and R~j are independently linear . S or branched C1-C6 alkyl groups, R4 is a linear or branched Ci-C12 alkyl or aryl group, and X Is a Cl-C~ alkyl group.
(iv) Polymers and copolymers of perfluoro-2,~-dimethyl-1,3-dioxole.
I A particularly useful polymer material falling within the general definitions above ; is poly~trimethylsilyipropyne) (PTM.SP~, which has the s~ructure:

! CH3 ~ {~` C = C }

`., . 10 : CH3 - Si - CH3 , ~ I

Mernbranes useful in the process of the inven~ion may be formed from these glassy, high-~rqe volum~ materials in a variety of ways. Because the materials are glassy and . rigid, an unsupported film of the polymer may often be usable as a single-layer gas .
, separation membrane. Altsrn2tively, the membrane may be an integral asymmetric 1~ membrane, comprising a dense regiorl that forms the se~aration membrane and a microporous support region. As a further alternative, the membrane may be a ~omposite iS
membrane comprising the separation membrane and an attached or unattached backing i layer, or a composite membrane comprising a microporous support membrane of one polymer coated with the separa-~on membrane of another polymer. Applicants prefer to ~j l;"t~
.~`, .
W0~4/0~86 ~li8~a9 PCl/VS93~10736 '.. ?j~; 1 4 use composite membranes.
h 'l The membrane incorporating the separation membrane may be îormed as a flat sheet, a hollow fiber or any other convenient form, and housed in any appropriate type of ;~;t module, such as plate-and-frame, potted fiber or spiral-wound.
In the process of the invention, a feed gas stream containing a condensable organic compound is passed through a membrane separation s~ep. The process is shown schematically in its very simplest form in Figure 1. Referring to this figure, an organic !:~
compound-containing feed gas stream, 1, passes to a membrane separation unit, 2,containirlg one or more membranes. The membrane separation step involves running the .~3~ 10 feed gas stream across a membrane that is selectively permeable to the organic compound that is to be remo~ ed. The organic compound, or, specifically in ~he case of natural gas treatment, the C~+ hydrocarbon fraction, is concentrated in ~he stream9 4, permeating the membrane; the residue, non-permeating; stream, 3, is correspondingly dep1eted in organic compound or higher hydrocarbons.
1~ If the f eed gas stream to be treated is at high pressure compared with atmospheric, the separation may be effected simply by making use of this high pressure ~o provide an adequate driving force and pressure ratio. Otherwise9 a pressure difference can be provided by compressing the feed stream, by drawing a vacuum on the permeate side of the membrane, or a combination of both. The pressure at which raw natural gas emerges from the well varies considerably from field to field, although pJessures up to 5000 psi are not uncommon. Polymer membranes can typically withstand pressure differences between the feed and permeate side up ~o abou~ 1,500-2000 psi, so, for this application, it may sometimes be necessary to lower the gas pressure before it can be fed to the membrane system. Also, depending on the origin and compositinn of the gas, it may be desirable to ¦ ~ 1 2~ I pass the gas through~a filter, phase separator, heater, e~c. before it enters ghe membrane system to remove entrained water or hydrocarbons in liquid or aerosol form. Any other pre-treatmene to remove contaminants or change the ~as co~nposition may also be performed as appropr-ate.
Single-stage gas-separation processes typically remove up to about 80-9~% of thepreferentially permeating component from the ~eed stream and produce a permeate stream that h~s five times or more the concentration of that component of the feed gas. This degree of separa~ion is adequate for many applications. If the resjdue stream requires ) WO 94J09886 2 1 4 8 ~ 3 ~ PCl/US93tlO736 ,.~3 15 further purification, it may be passed to a second bank of modules for a second processing step. If the permeate stream requires further concentration, it may be passed to a second bank of modules for a seeo~d-stage treatment. Such multistage or multistep processes, and varian~s thereof, are famil}ar to those of skill in the art, who will appreciate that the S p.ocess may be configured in many possible ways, including single-stage, multistage, multistep9 or more complicated arrays in series or cascade arrangements.
Optionally, the permeate stream from the membrane separation step may be recompressed and!or chilled to recover the organic compound in liquid form.
. Many pos~ible applications in the gas, oil, or petrochemical industries, for example, are envisaged. For this set of specific applieations of the invention, the cornposition of the .'''!3 feed gas may vary widely, from a mixture that contains 95%+ pure methane, with small amounts of ethane9 other hydrocarbons, water vapor, hydrogen sulfide, carbon dioxide and nitrogen, to streams that contain substantial percentages of C~ hydrocarbons or carbon dioxide. In many parts of the world, associated gases from oil wells are simply flared or reinjected. I~ is possible to use the membrane-based process described herein to remove propane and heavier components from associated gases, thereby producing natural gas liquids (NGL~ suitable for adding to the crude oil from the produ~tion separators. A
second possible application in the oil and gas industry is recovery of propane and higher hydrocarbons from: gas streams containing much carbon dioxide and/or nitrogen. Inert 2Q gases are used to recover incremental oil from partially depleted oil fields. When the oil is produced, large volumes of æsociated gases, which must be subjected to complex and costly processing steps if the valuable hydrocarbon component, are generated. The process of the invention offers a simple, economic treatment option. A third application in this area is in Btu control. The process of the invention may be used to remove and recover propane and heavier components and thereby reduce the Btu ~alue of pipeline or fuel gas.
A fourth application in natural gas treatment is hydrocarbon dewpoint control. Other applications include pretreatment of hydrocarbon-laden gas streams, to prevent damage to cellulose-based membranes, or other hydrocarbon-sensitive membranes, and NGI,recovery from refinery gases or off-gases from the petrochemical industry.
The invention is now further illustrated by th~ follQwiQg examples, which are intended to be illustrative of the invention, but are not intended to limit the seope or underlying prine,iples of the invention in any way.

:`.! WO 94/09886 i~ 1 ~ S t'J ~ 3 PCI/US93/10736 .,~ !

EXAMPLES
~i~ EXAMPLES 1-3. ComDarative exam~les with variou$ r~l~berv ~olvmers~ _ No~ in accordance with the inven~ion EXAMPLE I
;~, S An asymmetric, microporous polysulfone support membrane was prepared. A
sealing coa~ of silicone rubber about I ~m thick was applied to the skin side of the support membrane. Composite membranes were prepared by hand-coating a solution of ~ wt%
;, ethylene/vinyl acetate (E~rA, Elvax 150) in cyclohexane onto the skin side of the :.~
;~ asymme~ric support membrane by a continous dip-coating method. The polymer solution ~$ 10 was applied at room temperature and the composite membrane was air-dried.
i1 `:,! The membranes were tested with pure nitrogen, oxygen, and methane at 50 psig feed pressure to ensure that the EVA coating was defec;-free.
`~ The composite membranes with defect-free separation layers were evaluated in ~:
room-temperature gas-separation experiments. The membranes were mounted in a test ,.,.
ceil exposed to a gas mixture consisting of 86% methane, 10% ethane, 3% propane and 1%
butane on the feed side. The feed pressurc was maintained at 500 psig and the permeate side of the membrane was at atmospheric pressure. To maintain a constant feed gas composition, gas was continuously YeFIted from the high-pressure side to promote rnixing in the cellO The composit;ons of ~he residue and peFmeate strèams were analyzed with an 1 20 on-line gas chromatograph.
.~ The same preparation techn,que and permeation tests were repeated, except that the membrane was made from a 5 wt% Elvax 4~0 solution.
The pressure-normalized gas fluxes and the propane/methane and butane/methane selectivities of the membranes ar~ given in Table 1.
t ` ' I
Table 1. Permeation Properties of Elvax 150 and 450 Composite Membranes 'i .~ ~
~` Pressure-normalized fluxSelectivity ~x 10-6 cm~STP)/cm2 s cmHg~
i~ ~ _ _ MembranePureMethane in gas Propane/ l~utane/
methane mixturei me~hane methane `.~1 ' ~ __ . .~ _ _ ~ Elvax 1503.5 6.8 2.8 5.
___ ~ ~ __ ~_ Elvax 4501.4 5.6 1.9 3.0 1 ' ~ ~ ~

~ d ~ 2148~09 W~94fû9886 . PCr/US93/1073 Composite membranes were prepared by coating a solution of 3 wt% chlorina~ed polyethylene (25q~ chlorine grade) in 1,1 ,2-trichloroethane onto an asymmetric polyamide (Trogamid~ support membrane, using the same gene.al techniques as in Example 1.
Permeation tests were conducted as in Example 1.
,~ Composite membranes were prepared by coating a solution of 4 wt% ni~rile rubber (21% acrylonitrile) in methylethylketone (MEK) onto an asymmetric polyetherimide ~pFI) ; support membrane, using the same general techniques as in Example 1. The membrane was overcoated with a sealing layer of silicone rubber. Permeation tests were conducted as in Example l.
The pressure-normalized gas fluxes and the propane/methane and butane/methane selectivities of the membranes are given in Table 2.
Table 2. Permeation Propert}es of Polyethylene and Nitrile Rubber Membranes _ I~~n~ n"- Selectivity tx lo ~ cm~(STP)/cm2-s cmHg) _ ~ _ _ ~ . ...
1~ Membrane PureMethane in gas Propane/ Butane/
methane mixture methane meshane 1.; ~ . ~ . ._ ,1 Chlorinated 2.0 4.8 l .7 2.7 polyethylene . . , ~ _ _ ~ __ Nitrile 0.9. 2.1 1.6 2.8 rubber __ ~ __ Composite membranes were prepared by coating a solution of ~.7~ wt% silicone rubber in cyclohexane onto an asymmetric polysulfone support membrane, using the same general techniques as in Example l, but heating the membrane after coating to crosslink the silicone rubber. Permeation tests were conducted as in Example 1.
Permea~ion experiments were also carried out with some previously made composite membranes which consisted of a selec~ive layer of a polybutadiene/ silicone rubber blend on a polysulfone support.
l~he pressure-normalized gas fluxes and the propane/methane and butane/methane selectivities of the membranes are given in Table 3.

w~ g~/09~86 ~ 1 1 8 t~ ~1 9 PCI/US93/1û736 3~ T~ble 3. Permeation Properties of Silicone Rubber and Polybu~adiene Blend Membranes ~' " ,, . . . ... ..
Pressure-normalized flux Selectivity ~i (x lo-6 cm~(S ~P)/cm2 s crnHg) . _ Membrane Pure Methane in gas Propane/ Butane/
~' methane mixture methane methane .,.,~ ~ _ _ ....... _ _ .. , . -~i Silicone 4.2 S.8 3.7 6.1 rubber ~
~,. Polybutadiene 2.3 3 .0 5 .0 9.3 ii~3 blend ~ ~ ~
Reviewing Examples 1~3, it may be seen that in all cases, the methane flux measured with mixed gas samples is higher than the pure methane flux. This increase in methane flux is caused by plasticization of the membrane by the higher hydrocarbons in the feed. The higher methane flux resul~s in a lower hydrocarbon/methane selectivity.
., This behavior is typical of rubbery polmer membranes.

PTMSP films of thicknesses up to about 200~m were hand-cast from a solution of IS S% PTMSP in toluene onto gla~s plates, The films were mounted in a test cell and pure gas permeation measurements were made, using the same general ~echnique as in Example I, except that pure gases only were used and the feed gas pressure was ~0 psjg. The apparent thickness of the PTMSP layer was obtained by dividing the nitrogen permeability ¦ coefficient by the pure pressure-normali2ed ni~rogen flux through the membrane.
¦ 20 The ideal selectivity w~s calculated from the ratio of the pure gas pressure-~ normalized fluxes. The results are given in T~ble 4.
;
~.~

.

.

-~-~ WO94t098~6 ~1 4~ PCr/US93/10736 lable 4. Pure-Gas Transport Properties oî PTMSP

Gas 1'~ _ Selectivity Selectivity (Barr~r) (Gas/N ) (Gas/CH ) . _ . . . ~ .
Nitrogen 6,4û0 I .0 0.43 , _ _ Oxygen 9,600 1.5 0.6 _ . ~ , . ,,, ~
Methane 14,~00 2.3 1.0 `
_ , . ~ _ . . .. .;' Propane 40,000 6 3 llutane 94,000 14 6 6,4 These selectîvities appear to be too low to make this a practical method for separating more condensable from less c,ondensab1e components in gas mixeures.
EXAMPLE 5. omparison of Purç-~as ~nd mixed-gas behavior A set of experimen~s was performed to compare pure-gæ and mi%ed-gas fluxes, permeabilities and selectivities. In the first experiment, PTMSP films of ~hicknesses 16, 48, 90 and 200 ~m were prepared and tes~ed as in Example 4, ex~ept tl at the gases used for the testing were pure methane and pure butane. For the me~hane tests, the feed 1~ pressure was maintained at 50 psig and the permeate side of the ~embrane was at atmospheric pressure. For the butane tes~s, the feed p~essure was maintained at 5 psig and the perrneate side of the membrane was at atmospheric pressure. The pressure-normalized metb~ne and butane auxes are plotted as function of inverse PTMSP film thickness in Figure 3. As e~tpected, both plots are straight lines, with flux increasing in inverse proportion to membrane thic}cness.
lhe idoal butane/metbane selectivity was calculated from the flux daga. The -results are plotted in Figure 4. As can be seen from the figure, the ideal selectiYity remains esserltially constant at about; 5-S.~ over the thickness range. This selecti~fity does ~-not appear to be high enough ~o make this a practical method for separating bu~ane from , ~:
methane in gas mixtures.
Permeaeion tests were repeated with mixed gas sample~ co~sisting of 86% methane,10% ethane, 3% propane and I q6 butane. The ~eed pressure was maintained at 300 psig and the permeate side o~ the membrane was at atmospheric pressure. The pressslre-normalized -methane and butane fluxes are plotted as a function of inverse PTMSP film thickness in Figure ~. The methane plot appears to be still elose to a straight line, bu~ the butane flux .;

.
,. , .. ,.. - .. .. .. . .. ... ... , . . ~ .. . . ... ....... .. . . .

WO g4/09$~6 PCr/US93/10736 ~
~8~9 20 is clearly no longer in direct rela~ionship to the membrane thickness. The mixed gas selectivity is plotted as a function of the inverse PTMSP îilm thickness in Figure 6. lt is ~ery clear that the selectivity is now (i) better than the ideal selecsivity, and (ii) a function of membrane thickness. The maximum achievable butane~methane selectivity as measured at this feed pressure (300 psig) can be read from the ordinate of the graph and is about 50.
This is about lO times the ideal selecSivity obtained from the pure gas data.

Experimental and literature permeation data that we had gathered from a variety of rubbery, gla~sy and glassy, high-free-volume polymer materials was compared. A
comparison of representative materials is given in Table ~.
Tab!e 5. Comparison of ideal and mixed-gas selectivities for different types of polymer materials . . . ~ , . . . .
Ideal (Ppre-Gas) Mixed-Gas Membrane material Sele~ tivity Selec tivity Propane/ Butane/ Propane/ Butane/
methane methane methane methane ., _ . ~ ..
1~ Rubber Silicone rubber 7.3 33 3.8 6.~
. . ., _ _ . . .. . _ .

Polysulfone _ _ 0.14 0.51 0.60 _ _ _ . . . . , Hi~h-VFQ!~
PTMSP___ 2.6 5.6 8.2 27 :`
For the typical rubbery polymer, silicone rubber, the ideal selec~ivity for butane/methane is more than five times greater than the actual~ mixed-gas selectivity. Fo~
the typical glassy polymer, polysulfone, ~he polymer has an ideal butane/methane ~.
selectivity of 0.14, or a methane/butane selectivity of 7.1. The mixed-gas me~hane/butane 25! I selectivity is 1.7, so not only is she mater'al methane-seleetiYe, but again the ideal selectivity ;s higher, about four times higher ~han the actual mixed-gas selectivity. In con~rast, the PTMSP is butane-selective, and the actual, mixed-gas selectiYity is about five times greater than the ideal selectivity.

";:
. .:

.:

W(~ 94/09~6 2 1 ~ 8 ~ O ~ Pcr/US93/l~736 EXAMP~ E 7 An asymmetric, microporous support membrane was prepared. The support membrane, when te~sed with pure nitrogen, exhibited a pressure-normalized nitrogen flux ranging from 1.3-3.4 x ]o~l cm3(STP)/cm2-s-cmHg.
Composite membranes were prepared by coating a solution of 5~ polytri-methylsilylpropyne (PTMSPj in toluene onto the skin side of the asymmetric support membrane by a continous dip-coating method. The polymer solution was applied at a room temperature arld the composite membrane was dried in an oven at ~0-60C.
The membranes were tested with pure nitrogen, oxygen, and methane to ensure that the PTMSP coating was defect-free. The apparent thickness of the PTMSP layer was obtained by dividing the nitrogen permeability coefficient of PTMSP (- 6,400 Barrers) by the pure pressure-normalized nitrogen flux through the membrane.
A second set of membranes was prepared by performing the dip-coating step t-~vice to form a thicker separation membrane. A third set of membranes was prepared by lS repeating the dip-coating step three times. A fourth set of membranes was prepared by repeating the dip-coating step four times. The second, ~hird and fourth sets of membranes were tested with pure gases as above to check for defects. The apparent thickness of the separation membrane for ~he tWQ-, three- and four times-coated membranes was calculated in the same way as for the once-coated membrane.
The composite membranes with defect-free separation layers were evaluated in room-temperature gas-separation experiments. The membranes were mounted in a test cell exposed to a gas mixture consisting of 86% methane, 10% ethane, 3~6 prop~ne and 1%
butane on the feed side. The feed pressure was maintained at 100 psig and the permeate side of the membrane was at atmospheric pressure. To maintain a ~ons~an~ feed gas 25 i composigion, gas was continuously vented from the high-pressure side to,promote mixing in the cell. The compositions of the residue and permeate s~reams were analy~ed with an on-line gas chromatograph.
The apparent thickness of the PTMSP separation membrane, the pressure-normalized gas fluxes, and the propane/methane and butane/methane selectivities of the membranes are given in Table 6.

, .... ...... .. .

WO 94/09886 PCI'/U~93/10736 6 ~ 3 22 Table 6. Permeation Propereies of PTMSP Composite Membranes _ _ .__ ,, Separation Pressure-normalized flux Selectivity membrane ~ / c~ ~ _ _ thickness (~m) Pure me~hane Methane in gas Propane/ Butane/
mixture methane methane ~ _ . . .
7 (2 coats,~ __ 33 8 22 . . __ _ ~ ..... . .. , l I (3 coats~ _ 140 21 9 34 7 (~C~ ~ 96 _ 14 10_ 52_ As can be seen from Table 6, the separation membrane exhibited a marked selectivity/thickness relationship. The very high flux of the support membrane eliminates resistance of the underlying layer as a contributing factor to the selectivity.
EXAMPLE 8 ~ , PTMSP films of thickness 48 ~m and 200 ~m were hand-cast from a solution of 5 wt% PTMSP in toluene onto glass plates. Integrity and permeation ~ests were conducted ~`
as in E%ample 7. The gas mixture used for the permeation tests consisted of 86% methane, 10% ethane, 3% propane an~ 1% butane. The feed pressure was varied from 300 psig to 950 psig and the permeat~ side of the membrane was at a~mospheric pressure. The results ~ `
are given in Tables 7 and 8.

Table 7. Permeation Properties Qf 48-~m PTMSP Film at Different Feed Pressures . _ ~
Feed pressure Pressure-llormalized flux Selecti~ity ~
(psig) 1~ 1~ :~--~ ``
~Methane Propane Butano Propane/ Butane/
methane me~hane . . _ __ ~ _~ :
3Qo f ' ~ ~ 4 ~ 42 150 ~ g.8 ~ 31 ` ~
. .. _ . _ . . ... ~ :~ .
S00 S 5 41 ~ 10 7O~ ~0 .
600 6 l 42 ~6 6.9 1~ ~
, .. .. ~ _ ~ :.
800 5.8 35 - 72 6.0 ~2 -950 5 9 32 6l 5.4 lO -:..
: .
'.~'', :, ~ W~ ~4/09886 . 2 ~ 4 8 ~ 3 ~
23 ~
Tahie 8. Permeation Properties oî 200 ~m PTMSP Film at Different Feed Pressures ~ ~ . . ~
Feed pressure Pressure-normialized flux Selectivity (psig) ~ ~ :~_.........
Methane Propane Butane Propane/ Butane/
methane methane -:
__ _ . - ~ . .. . . ~
300 1.1 ~0 50 9.1 4 . _ ~ _ ~00 l.l 9.9 36 ~.û 33 ~_ ...................... , .. . , _ 600 1.2 9 9 31 8.2 ~6 . 800 1.4 9.5 26 6.8 19 . . ~ . -, . ~ _ _ 950 i 5 9 ~24 6.3 16 ~,, . ., . ,, ~ '":~
Tables 7 and 8 show that increasing the feed pressure }owers the selectivity. This ~.
drop in selectivity results from the sirnu?taneous increase of methane flux and decrease of .
propane and butane fluxes at hi8her pressures.

We prepared a 2-in diameter spiral-wound membrane module colltaining approxirnately 900 cm2 of a~tive membrane area. Ths membranes were composites l~ consisting of a PTMSP layer supported on a nonwoven po1yester fabric. The apparent PTMSP layer thickness was 45 ~ m.
The liydrocarbon separation properties of the modules were determined by permeation tests as in the previous examples, except that the module was mounted in the test system in place of the membra~le test eell. The module was exposed to a feed gas mi7cture consisting of 86% methane, 109~ ethane. 3% propane and 1% butane at feed pressures of 20û, 300, and 400 psig, the maximum pressure rating of the module housing.
The pressure-normalized gas fluxes and the propane/methane and bu~ane/methane selectivities of the modules are given in Table 9.

'.

.. . . ... , .. . ... ...... , .. ., . ~ .. . . . " ., . . . , ., . . . , .. . . .... , ., . . ~, ~, . . . . . .
. .

WO 94/0~86 P~r~VS9~/10736 Table 9. Permeation Properties of Lab-Scale PTMSP Membrane Module . , . . - . . . . . ~ .
Pressure-normalized flux Selectivity Feed pressure (x I o-6 cm ~ mHg) (psig)Methane Propane Butane Propane/ Butane/ -~
methane methane _ , =. ~- . .- . .. ~ , _ ..
200 3 9 31 ~ l~ 7.9 28 . ._ , _ . . . . , ., 300 3 6 27 85 7.5 24 _ _ _ _ .. _, . .
400 3.7 27 7~ 7.3 2l ~ '' The results show that the module effectively separated the higher hydrocarbons from methane.
, -, I ûThe preparation technique and permeation tests of Example 9 were repeated, using a module incorporating a membrane made from a higher molecular weight polymer. The ~ -results are given în Table 10.
,. .
Table 10. Permeation Properties of Lab-Scale Module with Higher MW Polymer , ~_ ~ _ ,~.
Pressure-normalized flux ~ Selectivity Feed pressure (x lO-~ cm~ ~STP)/cm2 ~ ~ _ (psig) Methane PropaneBut~ne Propane/ Butane/
methane methane ,, . ~ . . . . .
200 3 1 27 1~0 g.7 39 ~
., . . . ~ _ ~ : ':' 300 2.9 25 86 8.6 30 ;
~. ~ : . _ ... . ~ _ ,~, . ~. _ _~ ~ .. ,' ~00 2 8 ~2 62 7.8 22 ~: : ~ ~
Comparing the results of Examples 9 and 30 with Example 8, it may be seen that , the~module properties are comparable wi~h those~of the membrane. Between 200 and 400 psig feed pressures, the module selectivities are 7.2-8.7 for propane/methane and 21-39 for butanejmethano. Thoso Yalues are within the samo range as those obtained for a 48-~m PTMSP film atsimilar pre~sures.
~ExAMpLF ll ~ The preparation ~techni~ue and~ permeation tests of Example 10 were repeated, using a gas mixture consisting of 82% methane? 10% ethane, 7% propane and l% butane. -~
T he results are given in Table I I. - ;

- ..
:
.

21~8S~9 WO9~/098~6 PCr/US93/10736 Table 11. Permeation Properties of Lab-Scale Module with Higher MW Polymer ~ ,............... , ,__ . ~ _ Pressure-normali7ed flux SelectiYity Feed pressure (x 10~~ cm ~ ~
(psig) Methane Propane Butane Propane/ Butane/
methane methane , _. . _ . . ,__ _ 200 3.6 30 120 B.3 33 _ _ __ __ . __ __ . ;. .
300 2.~ 22 72 7.6 25 _ _ . . _ _ _ .~ __ ~
400 2 9 21 61 ~.2 21 __ _ __ . ~_ __ ' Once again, the results were similar to those achieved with membrane stamps of the same thiekness.

lOComposite membranes were prepared by dip-coating solutions of 5% polytri- ;
methylsilylpropyne (PTMSP) in toluene onto a nonwo~en polyester fabric backing. The resulting composite membranes were tes~ed with pure nitrogen, oxygen, and methane to ensure that the PTMSP coating was defect free. The apparent thickness of the PTMSP
layer was obtained by dividing the nitrogen permeability coefficient of PT~SP (~ 6,4û0 15Barrers) by the pure pressure-normalized nitrogen flu~ ~hrouph the membrane. The fJuxes were consistent with a membrane ~hickness of about 30 ~m thiclc.
The composite membranes with defect-free separation layers were evaluated in ~`
room-temperaturo gas-separation experiments. The membranes were mounted in a test celi exposed to a gas mixture consisting of 8~% methane, 10% ethane, 3% propane and 1%
20butane ~n the feed side. The feed pressure was mainta;ned a~ 500 psig and the permeate side of the membrane was at atmospheric pressure. To maintaila a constant feed gas composition, ~2S was continuously vented from the high-pressure side to promote m;xing in the cell. The compositions of the residue and permeatc streams wer~ analyzed with an on-line gas chromatograph.
2SThe permeation tests wsre repeated with a gas mixture saturated with hexane.
Saturation was obtained at about 600 psig and 25C ~y passing the pressurized gas mixture as above throug~. a high-pressure bubbler containing liquid hexane. The resul~ing feed composition was 85.7% me~hane, 9.6% ethane, 2.9% propane~ 1.0% butane, and 0.8%
- hexane.

;-,.

WO 94iO9~86 ~ 6 D ~ PCT/US93/10736 The selectivity for the various compunents over methane, in mixtures with and without hexane, are given in Table 12.
Ta~le 12. Selectivi~y of Pl MSP Membrane with and without n-Hexane in Gas Mixture Selecti-rity __ lle~ e s~ t~ d Feed mix~wi~hout hexane S Ethane/methane 3 .1 3 .0 _ . . . , . _ Propane/methane 6.4 6.2 , ... .. _. .
Butane/msthane 14.5 15 .0 H~x~o~/m~tllan~ 48.7 ___ The results shown in Table 12 demonstrate that the butane/me~hane separation performance of the membrane was not affected by the presence of hexane in the feed.
The selectivities obtained with the hexane-saturated feed were 3.1 for ethane/methane9 6.4 for propane/methane, and 14.5 for butane/methane. With the feed containing no hexane, the corresponding selectivities were 3.0, 6.2, and 15Ø These two sets of values ~iffer by only 3 to 4%, well within the range of experimental error. The data also indicate that the PTMSP membrane permeated hexane 49 times faster ~han methane, leadin3 to e~cellent hexane recovery.
EXAMPLlE 13 ; The experiment of Example 12 was repeated~ except that this time a less-condensable gas, ~arbon dioxide, was added, so that the feed gas mixture had a composition of 86% methane, 10% carbon dioxide, 3% propane and 1% butane. The permeation tests were carried out using a spiral-wound module, as in Examples 9 and 10.
The feed gas ~ressure was 300 psig; the permeate side of the module was at atmospheric pressure. The results are listed in Table 13.
I I ~ ' , j j ., I` ~ ,.. `:

. .-:..
"

~ ' .. ~

~ ~ W094/û98~6 i~2:1 18~,.9~ Pcr/~sg3/lo736 Table 13. Selectivity of PTMSP Membrane with and without Carbon Dioxide in Gas Mixture ~c~iv~_ C0~-cont ining feed _ Feed mixture wlthout C02 Ethane/methane _ _ 3 .3 _ .
Propane~methane 7.8 8 .0 Butane/methane 25 .8 2~ .7 , , , . . ~ -- . ~ ._ _ C0./~lb~ -No significant change in the separation properties of the membrane module was found when carbon dioxide was present in the feed. The selectivities were 7.3 for propane/methane and 25.8 for butane/methane, equivalent to the values of 8.0 and 25.7 obtained with the gas mixture without carbon dioxide. Furthermore, the mixed-gas carbon dioxide/methane selectivity of 2.8 is abDut 30% higher than the pure-gas value of 2.1. As expected, the pressure-normalized mixed-gas methane flux was more than ~ times smaller than the pure-gas methane flux because of the co-permeation of the larger, more condensable hydrocarbons. A similar reduc~ion was observed for the mi%ed-gas carbon dioxide flux, compared ~o the pure-gas carbon dioxide flux. Thust the presence of less condensable sp~cies in the feed stream did not affect the separation perfonnance of the membrane to any significant extent.

. .

Claims (25)

We claim:
1. A process for recovering a condensable organic component from a gas stream, comprising the steps of:

(a) providing an incoming gas stream containing a condensable organic component, said condensable organic component being characterized by a boiling point higher than -50°C
at atmospheric pressure;

(b) providing a membrane having a feed side and a permeate side;

said membrane comprising a polymer material characterized by:

(i) a glass transition temperature, Tg, of at least about 100°C, (ii) a free volume VF of at least about 10%;

(c) contacting said feed side with said gas stream;

(d) withdrawing from said permeate side a permeate stream enriched in said condensable organic component compared with said gas stream:

(e) withdrawing from said feed side a residue stream depleted in said condensable organic component compared with said gas stream.
2. The process of claim 1, wherein said glass transition temperature, Tg, is at least about 200°C.
3. The process of claim 1, wherein said free volume VF is at least about 20%.
4. The process of claim 1, wherein said polymer material is a substituted polyacetylene.
5. The process of claim 1, wherein said polymer material is a silicon-containingpolyacetylene.
6. The process of claim 1, wherein said polymer material is a germanium-containing polyacetylene.
7. The process of claim 1, wherein said polymer is polytrimethylsilylpropyne.
8. The process of claim 1, wherein said polymer is poly(perfluoro-2,2-dimethyl-1,3-dioxole).
9. The process of claim 1, wherein said polymer material has an ideal selectivity for said condensable organic component over a second component of said gas mixture and a mixed-gas selectivity for said condensable organic component over a second component of said gas mixture and said mixed gas selectivity is greater than said ideal selectivity.
10. The process of claim 9, wherein said mixed gas selectivity is at least five times greater than said ideal selectivity.
11. The process of claim 9, wherein said mixed gas selectivity is at least ten times greater than said ideal selectivity.
12. The process of claim 1, wherein said gas mixture comprises an organic vapor in air.
13. The process of claim 1, wherein said gas mixture comprises an organic vapor in nitrogen.
14. The process of claim 1, wherein said gas mixture comprises an organic vapor in hydrogen.
15. The process of claim 1, wherein said gas mixture comprises an organic vapor i methane.
16. The process of claim 1, wherein said condensable organic component comprises a compound selected from the group consisting of aliphatic hydrocarbons, aromatic hydrocarbons and halogenated hydrocarbons.
17. The process of claim 1, wherein said gas mixture comprises natural gas.
18. A process for separating a C3+ hydrocarbon from a natural gas stream, comprising the steps of:

(a) providing an incoming natural gas stream;
(b) providing a poly(trimethylsilylpropyne) membrane having a feed side and a permeate side;

(c) contacting said feed side with said gas stream;

(d) withdrawing from said permeate side a permeate stream enriched in said C3+
hydrocarbon compared with said natural gas stream:

(e) withdrawing from said feed side a residue stream depleted in said C3+ hydrocarbon compared with said natural gas stream.
19. The process of claim 18, wherein said C3+ hydrocarbon comprises propane.
20. The process of claim 18, wherein said C3+ hydrocarbon comprises butane.
21. The process of claim 18, wherein said membrane exhibits a mixed gas selectivity for butane over methane of at least 12.
22. The process of claim 18, wherein said membrane exhibits a mixed gas selectivity for propane over methane of at least 6.
23. The process of claim 1 or claim 18, wherein said membrane is a composite membrane.
24. The process of claim 1 or claim 18, wherein said membrane is an asymmetric membrane.
25. The process of claim 1 or claim 18, further comprising condensing at least a portion of said permeate stream.
CA002148609A 1992-11-04 1993-11-03 Gas separation process Abandoned CA2148609A1 (en)

Applications Claiming Priority (4)

Application Number Priority Date Filing Date Title
US97069992A 1992-11-04 1992-11-04
US07/971,331 1992-11-04
US07/971,331 US5281255A (en) 1992-11-04 1992-11-04 Gas-separation process
US07/970,699 1992-11-04

Publications (1)

Publication Number Publication Date
CA2148609A1 true CA2148609A1 (en) 1994-05-11

Family

ID=27130538

Family Applications (1)

Application Number Title Priority Date Filing Date
CA002148609A Abandoned CA2148609A1 (en) 1992-11-04 1993-11-03 Gas separation process

Country Status (4)

Country Link
EP (1) EP0667803A4 (en)
JP (1) JPH08502926A (en)
CA (1) CA2148609A1 (en)
WO (1) WO1994009886A1 (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10293301B2 (en) * 2017-02-09 2019-05-21 Saudi Arabian Oil Company Modified siloxane composite membranes for heavy hydrocarbon recovery

Family Cites Families (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPS59154106A (en) * 1983-02-23 1984-09-03 Toshinobu Higashimura Gas-separation membrane
US4657564A (en) * 1985-12-13 1987-04-14 Air Products And Chemicals, Inc. Fluorinated polymeric membranes for gas separation processes
US4759776A (en) * 1986-12-08 1988-07-26 Air Products And Chemicals, Inc. Polytrialkylgermylpropyne polymers and membranes
DE3806107C2 (en) * 1988-02-26 1994-06-23 Geesthacht Gkss Forschung Process for discharging organic compounds from air / permanent gas mixtures
US4859215A (en) * 1988-05-02 1989-08-22 Air Products And Chemicals, Inc. Polymeric membrane for gas separation
US4931181A (en) * 1989-06-02 1990-06-05 Membrane Technology & Research, Inc. Composite membranes for fluid separations
US5051114B2 (en) * 1989-06-15 1996-01-16 Du Pont Canada Perfluorodioxole membranes
US5013338A (en) * 1989-09-01 1991-05-07 Air Products And Chemicals, Inc. Plasma-assisted polymerization of monomers onto polymers and gas separation membranes produced thereby

Also Published As

Publication number Publication date
WO1994009886A1 (en) 1994-05-11
JPH08502926A (en) 1996-04-02
EP0667803A1 (en) 1995-08-23
EP0667803A4 (en) 1997-04-02

Similar Documents

Publication Publication Date Title
US5281255A (en) Gas-separation process
US5501722A (en) Natural gas treatment process using PTMSP membrane
Pinnau et al. Transport of organic vapors through poly (1-trimethylsilyl-1-propyne)
Pinnau et al. Hydrocarbon/hydrogen mixed gas permeation in poly (1‐trimethylsilyl‐1‐propyne)(PTMSP), poly (1‐phenyl‐1‐propyne)(PPP), and PTMSP/PPP blends
US4857078A (en) Process for separating higher hydrocarbons from natural or produced gas streams
Krishna et al. The Maxwell–Stefan description of mass transport across zeolite
Rao et al. Nanoporous carbon membranes for separation of gas mixtures by selective surface flow
EP0502877B1 (en) Process for recovering condensable components from gas streams
Pinnau et al. Gas and vapor transport properties of amorphous perfluorinated copolymer membranes based on 2, 2-bistrifluoromethyl-4, 5-difluoro-1, 3-dioxole/tetrafluoroethylene
Jiang et al. Performance of silicone-coated polymeric membrane in separation of hydrocarbons and nitrogen mixtures
Pinnau et al. Pure-and mixed-gas permeation properties of polydimethylsiloxane for hydrocarbon/methane and hydrocarbon/hydrogen separation
US5032148A (en) Membrane fractionation process
Leemann et al. Vapour permeation for the recovery of organic solvents from waste air streams: separation capacities and process optimization
US6572679B2 (en) Gas separation using organic-vapor-resistant membranes in conjunction with organic-vapor-selective membranes
AU2008358898B2 (en) Mixed matrix membranes incorporating microporous polymers as fillers
EP1378285B1 (en) Gas separation using membranes formed from blends of perfluorinated polymers
US20070135670A1 (en) Method of Separating Olefins from Mixtures With Paraffins
US6572680B2 (en) Carbon dioxide gas separation using organic-vapor-resistant membranes
Shi et al. Poly (dimethyl siloxane) thin film composite membranes for propylene separation from nitrogen
WO1997032171A1 (en) Membrane-augmented cryogenic methane/nitrogen separation
WO2018093487A1 (en) High flux, cross-linked, fumed silica reinforced polyorganosiloxane membranes for separations
CA1294093C (en) Polyphosphazene fluid separation membranes
US5707423A (en) Substituted polyacetylene separation membrane
US5688307A (en) Separation of low-boiling gases using super-glassy membranes
Nitsche et al. Separation of organic vapors by means of membranes

Legal Events

Date Code Title Description
FZDE Discontinued