'G
`
`.. VOLUME 113, NO.2
`... (Completing Vol. 113)
`'i
`
`15 MAY 1996
`
`ISSN: 0376-7388
`
`·ournalof
`EMBRANE
`SCIENCE
`
`UNIVERSITY Of WASHINGTON
`
`J UN 14 1996
`
`tlBRARlfS
`
`Special Issue
`Selected papers presented at the
`Second International Symposium
`Progress in Membrane Science and Technology
`Enschede, The Netherlands, 27 June-1 July 1994
`
`ELSEVIER
`
`

`
`ELSEVIER
`
`Journ al of Membrane Science 113 ( 1996 ) 275-284
`
`journal of
`MEMBRANE
`SCIENCE
`
`U e of nonporous polymeric flat-sheet gas-separation membranes
`in a membrane-liquid contactor: experimental studies
`
`D.G. Bessarabov "". E.P. Jacobs a, R.D. Sanderson a, l.N. Beckman b
`• Unive rsity ofStellenbosch. Institute fo r Polymer Science. Stellenbosch. 7600. South Africa
`" Moscow State University. Department of Chemistry. Moscow. 119899. GSP3. Russian Federation
`
`Accept ed 3 April 1995
`
`Abstract
`
`Rat-sheet non-porous asymmetric poly ( vinyltrimethylsilane ) ( PYTMS) membrane s and compo site membran es comprising
`J dense layer of polydimethylsiloxane/polyphenylsilsesquioxane ( PDMS/ PPSQ) block copolymer were evaluated for low(cid:173)
`temperature bubble-free deoxygenation of water flowing in a two-channel countercurrent liquid-membrane contactor. A novel
`arge-scale three-channel flowing-liquid -membrane module (se lective membran e valve ) . designed for gas separation, is also
`described. The system compri sed PYTMS or PDMS/PPSQ non-porous membranes which acted as gas-permeabl e barriers. A
`membrane system ( SMY) in which pure water formed a flowing liquid membrane was evaluated to control hydrogen transfer
`rate. The liquid flowing along the turbulence-prom oter spacers between the membranes reduced liquid-film resistance. The
`overall mass-transfer coefficient s were found to be a function of the liquid flow rate. The liquid-film resistance controlled the
`rate ofgas transfer in such membrane contactors.
`
`Kevwords: Gas . epara tions: Liqu id membranes; Membrane-liquid co ntactors: Flowin g liquid membranes
`
`1. Intr od uctio n
`
`Numerous concepts in the design and performance
`of liquid-membrane sys tems for gas separation and
`purification have been proposed; the se include: immo(cid:173)
`bilized liquid membranes (ILM)
`[1-5]; membrane
`[6-12] ; membrane permabsorbers
`contactors ( MC)
`(~IP) [13-19]
`( these contain non-porous gas-sepa(cid:173)
`ration membranes); flowing liquid membranes ( FLM)
`[20-25] and hollow-fiber-contained liquid membranes
`HFCLM) [26-29] . For a variety of reasons, accord-
`II1g to Kimura and Walment [30], liquid membranes
`which are immobi lized inside the pores of polymeric
`upports are not stab le.
`
`• Corresponding author.
`
`In the case of ga s separation by means of membrane(cid:173)
`contacting sys te ms ( M e. MP, HFCLM , etc.), poly(cid:173)
`meric membranes (porous or non -porous ) serve as the
`contacting media. As the polymeric membrane is either
`porous and hydrophobic or non -porous, the liquid and
`gas phases are separated from each other, making it
`possible for gas absorption by means of membranes to
`take place. Liquid-membrane contacting sy stems com(cid:173)
`bine absorption selectivity with the technical advan(cid:173)
`tages of membrane processes. The main advantages of
`the contacting liquid-membrane systems for ga s se pa(cid:173)
`ration have been summarized elsewhere [6-19,31,32] .
`Earlier studies on membrane gas-liquid contactors
`focused on systems operating with porous membranes
`in the non-wetted mode ( gas -filled pores ) and in the
`wetted mode ( absorp tio n liquid-filled pores ) [ 12,26-
`
`376-7388/ 96/ $15.00 © 1996 Elsevier Scien ce B.Y. Allrights reserved
`~Dl 0 3 7 6 - 7 3 8 8 ( 9 5 ) 0 01 2 6 - 3
`
`

`
`276
`
`D.G. Bessarab ov et 01. / Journal of Membrane Science 113 (1996) 275-284
`
`29]. It has been shown that the liquid-film resistance
`controls the mass transfer in gas-separation liquid(cid:173)
`membrane systems if the gas component does not react
`chemically with the liquid-phase [8, II].
`The development of thin gas-permeable polymeric
`membranes has stimulated research to evaluate their
`utilization in liquid-membrane contactors. It has been
`that the presence of a thin and highly
`shown [11 ,18]
`permeable non-porous film, interposed between the gas
`and liquid phases, does not strongly affect the overall
`mass-tran sfer resistance to gas permeation. The advan(cid:173)
`tages of this technique were summarized by Bessarabov
`et al. [ 18] who demonstrated also the main disadvan(cid:173)
`tage of such a system, that is, pollution of the gas phase
`because of the pervaporation ofa hot solution ofa liquid
`absorbent through the non -porous membrane.
`Where MC was used as a liquid-phase oxygenator,
`microporous hydrophobic hollow fibers of polypropyl(cid:173)
`ene were applied to oxygenate water [33], and the
`performance of a porous-membrane water degasser has
`been dem on strated [34,35] . Bubble-free aeration of
`water has been achieved by means of membranes, and
`a mass-transfer analysis has been presented [36]. In
`these studies two types of membranes which show use(cid:173)
`ful oxyge n-transfer properties have been identified:
`microporous hydrophobic polymeric membranes and
`thin silico ne rubber films [37].
`In this paper, emphasis is placed on the design of
`flowing-liquid membrane systems comprising non(cid:173)
`porous polymeric flat membranes for gas separation,
`and on the evaluation of such systems in various oper(cid:173)
`ations. It is discussed how the modification of the two (cid:173)
`channel membrane-liquid contactor, which has been
`used to deoxygenate water, enables to apply it to gas
`separation.
`
`2. Experimental
`
`Tw o type s of the membrane contacting modules
`have been used in this study: a two-channel membrane
`
`Table I
`Perm eabiliti es of oxygen in vari ou s polymers
`
`Pol ymer
`
`Oxygen permeability
`(c m '( ST P) em/e m's cmHg )
`
`6.0 X lO-x
`PDMS
`PDM S /PPSQ 4 .5 X lO- x
`PVTMS
`3.6 X 1O- '}
`
`References
`
`[42)
`Thi s study
`(42J
`
`contactor (MC) and three-channel selective membrane
`valve (SMV ). A detailed description of the large-scale
`MC module, used in absorption-desorption modes and
`comprising the PDMS/PPSQ composite gas-separa(cid:173)
`tion membrane, was given in an earlier paper describing
`the separation of an ethylene and ethane gas mixture
`[ 18].
`
`2.1. Deoxygenation ofwater flowing in a two(cid:173)
`channel membrane-liquid contactor
`
`A two-channel membrane cassette and stainless steel
`frame (MC), as described earlier [18], was used in
`this study. The multimembrane cassette comprised 24
`asymmetric membranes made from PVTMS or POMSI
`PPSQ-coated composite membranes. PVTMS asym(cid:173)
`metric
`and
`PDMS/PPSQ-coated
`composite
`membranes were used in the deoxygenating experi(cid:173)
`ments. The use of asymmetric gas separation mem(cid:173)
`branes offers high productivity because of the presence
`of a very thin diffusion layer in such membranes. Table
`I shows the permeabilities of POMS, PVTMS and
`POMS/PPSQ block copolymer. Table 2 shows the
`oxygen permeance ( PI I ) of the asymmetric mem(cid:173)
`branes used in this study and of the silicone rubber film
`used in previous work [37]. The SEM micrograph
`illustrates the structure of a PVTMS asymmetric
`membrane (Fig. I). A schematic diagram of the deox(cid:173)
`ygenation of flowing water is shown in Fig. 2.
`Distilled water with a conductivity of 4.0 I /-LS Iem
`was used. A peristaltic pump (Masterflex pump con(cid:173)
`troller, Cole-Parmer Instrument Co.) fed water into the
`system in a once-through mode (flowing mode) and
`the flow rates of gases and liquids were measured by
`rotameters. The oxygen concentration in the water was
`determined by means of a digital oxygen meter (Scott
`- 1050 mbar was
`Gerate, CG 867). A vacuum of
`applied to the gas channels in the module. The initial
`oxygen concentration in water was 8.6 mg/l. The tem(cid:173)
`perature of the water was 2 I-22°C throughout. Poly(cid:173)
`meric net-like turbulence promoters similar to those
`[18] were used to
`described by Bessarabov et al.
`enhance the mass transfer. The hydraulic diameter, dh,
`of the membrane module was defined by means of the
`equation derived by Schock and Miquel [39] as:
`
`4e
`21h + ( I - e)A v.sp
`
`=0.031
`
`(I)
`
`

`
`D.C . Bessarab ov et al. / Journal of Membrane Science //3 (/ 996) 275-284
`
`277
`
`~able2
`) ygen permeances of various polymeric membranes
`
`Membrane thickness (em)
`
`Oxygen permeance (cm3( STP)/cm> s cmHg)
`
`Remarks
`
`1.5 X 10- >
`""5.5x 10- 4
`""2.3 X 10- 4
`
`•
`
`•
`
`4.0 X 10- 6
`8.2 X 1O - 5 b
`1.9x 1O - 5 b
`
`Nonporous film 1371
`Composite membrane
`Asymmetric membrane
`
`Thickness of the dense diffusional layer.
`'Gas-membrane-gas" experiments to determine the permeances of these membranes showed variations in perrneances. This divergence in
`he data isexplained by the non-uniform thickness of the membrane dense layer.
`
`All geometric variables of the spacer were determined
`by means of visual measurements with a microscope.
`
`2.2. Large-scale three-channel flowin g-liquid
`membrane contactor
`
`In this study a description of a novel three-channel
`flowing-liquid-membrane module for ternary gas sep(cid:173)
`aration, that is, a selective membrane valve (SMV) is
`presented. In an SMV a liquid flows between two non (cid:173)
`porous membranes (Fig. 3), and a ternary gas mixture
`can be separated into its components by the membrane
`valve:
`the first component which is insoluble in the
`liquid membrane phase,
`the second which diffuses
`through the membrane sandwich, and the third com(cid:173)
`ponent which reacts chemically with the
`liquid
`membrane phase and is pumped to a desorption module
`for degassing. The selective membrane valve therefore
`has one inlet for the feed and three outlets for the
`products.
`To obtain the highest efficiency from the SMV with
`the minimum of equipment a considerably modified
`two -channel multimembrane module, described earlier
`[ 18], was used. The key element of the gas -separation
`system is the membrane cassette (Fig. 4). The modi (cid:173)
`fication comprised addition of a third independent
`these
`into the cassette. Fig. 4 shows that
`channel
`membrane cassettes have three branch ports (one for
`the entry of feed-gas; one for the entry of liquid; and
`one for the entry of sweep-gas). Fig. 5 shows one of
`these membrane cassettes placed on the surface of the
`stainless steel block. The membrane cas sette is com(cid:173)
`pressed between two blocks of stainless steel. Fig. 6
`shows three cassettes between two blocks. The bottom
`block has two branch pipes (one for the entry and one
`for exit of gas or liquid) and nine ports, through which
`gas (or liquid) is fed to contact the membranes. The
`
`Fig. I. Cross-section of the PVTMS asymmetric membrane (view
`[the porous sublayer).
`
`Spacers
`
`~~~
`
`Water How
`
`~.
`•
`
`••
`
`Vacuum
`
`Fig. 2. Schematic diagram of the deoxygenation of water flowing in
`atwo-channel membrane contactor.
`
`where Eis porosity of the spacer andA v.spis the specific
`urface area of the spacer [39] :
`
`(2)
`
`The value for porosity was obtained from its definition
`[39] :
`
`(3)
`
`

`
`278
`
`D.G. Bessarabov et al. I Journal of Membrane Science //3 (/996) 275-284
`
`upper block has three branch pipes (one each for feed(cid:173)
`gas, for sweep-gas and for liquid). When the upper
`pipes are locked, the membrane system can operate as
`a conventional MP (Fig. 7). The same polymeric net(cid:173)
`like spacers were used to enhance mass transfer, as
`descri bed by Bessarabov et al. [ 18] for the two-channel
`module. The space between the membranes was 0,045
`em. The total membrane area in one cassette was 0.57
`m2
`
`ig. 8 shows clearly the principle of the performance
`of this cassette. Although this diagram shows only three
`membranes between the steel blocks, the real cassette
`contains 24 flat-sheet asymmetric membranes made
`
`. F
`
`Permeate+
`
`Fig. 3. Schematic diagram of the three-channel membrane " sand(cid:173)
`wich ": the middle layer is a flowing liquid membrane.
`
`Fig . 4 . Three-channel multimembrane cassettes.
`
`Fig. 6. A three-channel liquid-membrane contacting system for gal
`separation. The system contains three casettes.
`
`Fig . 5. The bottom block of the SMV and a three-channel cassette
`on it.
`
`Fig. 7. A three-channel
`locked upper ports .
`
`liquid-membrane contacting system with
`
`

`
`D.G. Bessarabov et al. / Journal ofMembrane Science 1/ 3 ( /996) 275- 284
`
`279
`
`Permeate
`
`Block of stainless steel
`
`Sweep-gas
`channel
`
`. . . Retentate
`
`+-
`
`Water flow
`
`Spacers
`
`Block of stainless steel
`
`Fig. 8. Schematic diagram of the multimembrane three-channel cassette along with a stainless steel frame.
`
`trom PVTMS [ 17 ) or PDMS /PPSQ co mpos ite mem(cid:173)
`branes [ 18 J.
`
`~J. Hydrogen permea tion throu gh the SMV
`imprising PVTMS membranes in the flowin g mode
`
`The purpose of this study was to dem onstrate a new
`, ign for a Iiquid-mernbrane co ntacting system for gas
`eparation. A peristaltic pump was used to feed water
`Intothesystem . The flow rate s of water were measured
`h~ flow-meter s. In this study the single SMV module,
`perating in the flowing mode, was used. The feed-gas
`lias technical-grade hyd rogen ,
`and
`the
`flowing
`membrane was distilled water. Hydrogen was fed into
`ihe ystcm through the upper port as shown in Fig . 6.
`The hydrogen co nce ntration in the perm eate was deter-
`
`Tabid
`Experimental
`ripping
`
`results for water deoxygenation using vacuum
`
`aler flow rate
`ml s)
`
`Concentration of oxygen in water (mg/I)
`
`PYTMS
`membranes
`
`PDMS/PPSQ
`membranes
`
`2.3
`2.0
`1.2
`0.7
`0.6
`0.6
`0.6
`I
`
`:0
`
`2.3
`2.0
`0.9
`0.6
`0.6
`0.5
`0.5
`I
`
`mined with a Varian gas chromatograph ( model 3700 ),
`at room temperature. The carrier gas and the sweep gas
`were nitrogen. A TeD was used. The flow rate of the
`sweep gas was 20 mI l s. The separation data were
`recorded by a Hewlett Packard ( 3380A)
`inte grat or.
`The steady-state productivity, that is, (perrneance, p i
`I, crrr' / crrr' s cmHg ) was calculated using the same
`technique as was described in a pre vious paper [ 18).
`Since nitrogen was used as a sweep gas on the permeate
`side of the SMV as a model replacement for vacuum
`stripping, the exiting hydroge n stream contained nitro(cid:173)
`gen. The presence of nitrogen and water vapor on the
`permeate side of the SMV was not taken into acc ount
`in our calculation of hydrogen permeance.
`
`3. Results and discussion
`
`3. /. Deoxygenating of water
`
`Table 3 shows the oxygen concentration obtained at
`various water-flow rates. The overall mass tra nsfer
`coefficient ( Ka v ) for oxyge n removal can be calcu(cid:173)
`lated by the following equation [3 8] :
`
`Ka v =.Qln( [02 ]in)
`[02Lu
`aL
`Where [°2 Lnand [°2 LUI are the inlet and outlet oxy (cid:173)
`gen concentrations in the liquid, respectively, a is the
`membrane area per volume of the membrane bed
`( a = 31.7 em - I), L is the length of the membranes in
`
`(4)
`
`

`
`280
`
`D.C . Bessarabov et al. / Journal of Membrane Science 113 (1996) 275-284
`
`Table 4
`Dependence of the oxygen overall mass transfe r coefficient and the liquid film resistance on the liquid flow rate in the deoxyge nator. operating
`in the once-through mode
`
`Superficial water velocity ( crn/ s )
`
`Ove rall mass transfer coe fficie nt Kov X 10' (c m/s)
`
`Resistance liquid /t otal ( %)
`
`PVTMS membrane
`
`PDMS /PPSQ membrane
`
`PVTMS membrane
`
`PDMS /PPSQ membrane
`
`0.06
`0. 12
`0.25
`0.5
`0.62
`1.23
`1.73
`2.57
`
`1.3
`2.8
`7.8
`19.9
`26. 1
`5 1.8
`72.9
`87.5
`
`1.3
`2.8
`8.9
`21.1
`26. 1
`55.4
`77.9
`87.5
`
`99.7
`99 .3
`98.2
`95.3
`93 .9
`87.8
`82.9
`79.4
`
`99 .9
`99.8
`99.5
`99.0
`98 .8
`97.5
`96 .5
`96 .1
`
`the directi on of flow ( L = 20 cm ), and n is the super(cid:173)
`liquid ve loc ity in the module, that is, the volu(cid:173)
`ficial
`metric flow rate divided by cross-section for the flow .
`For the de oxygenation experiments the overall tran s(cid:173)
`port resistance ( I I Kov ) is gi ven by the sum of the
`membrane resi stance (II PH ) and the resistance of the
`boundary liquid layer ( 1/K L ) [36]:
`
`I
`I
`I
`-
`= - +(cid:173)
`Ko v PH K L
`
`(5)
`
`wher e H is Henry ' s co nstant, P is the polym er perme(cid:173)
`ability, and I is the thi ckness of the membrane den se
`layer.
`Table 4 sho ws the dependence of the overa ll mass
`transfer coefficient and liquid film resi stance on the
`
`100.--
`
`-
`
`-
`
`-
`
`-
`
`-
`
`-
`
`-
`
`-
`
`-
`
`-
`
`-
`
`-
`
`-
`
`-
`
`----,
`
`liquid flow rate. Fig. 9 shows the relationship between
`the mass transfer coefficient and water velocity though
`the following correlation [39] :
`
`(6)
`
`where Sh = Kovhdl D; Re = hdfll v; Sc = v i D.
`The div ergen ce in co rrelation between ex perimental
`and calculated results can be explained in the light of
`the sensitivity of the mass tran sfer coeffic ient for the
`prevailing hydraulic conditions when turbul enc e-pro(cid:173)
`moter spacers wer e used. It is clear from Table 4 and
`Fig. 9 that the liquid-film resistan ce co ntrols oxygen
`tran sfer in such a sy stem . As the liquid flow rate
`the liquid-boundary res ista nce decreases
`increases,
`slightly. Th e experimental data sho w that the ind ividual
`selectivities o f the non-porous gas-separation poly(cid:173)
`are
`not
`imp ortant
`in such a
`meric membranes
`membrane contacting system. For example, selectivity
`( a ) of the PVTMS asymmetric membrane for oxygen
`and nitrogen ( a 0 2/N2) is at lea st 3.5 and selectivity
`of a PDMS/PPSQ composite membrane for oxygen
`and nitrogen is at least 2.1. Membrane selec tivity may
`be important at very high liquid flow rates. In order to
`obtain such a high liquid flow rate inside a module. it
`is nece ssary to apply a rather high pre ssure at the inlet
`of the liquid channels; this is not of practical intere t
`here . Where there is a fast chemical reacti on between
`gas and liquid, the selectivity of the separation process
`is determined by selec tive absorption. Once again
`important.
`the membranes are not
`se lectivities of
`Hen ce, the membranes for such a sys tem should be
`se lec ted on the basis of their mech an ical propert ies. It
`is possible, therefore, to predi ct that the "fast" silicon
`
`10
`
`0.1
`
`0 .01
`0.1
`
`""
`
`"
`
`1
`Reynolds Number
`
`10
`
`}j
`
`E:
`
`J
`
`Z8~'
`
`"J:;en
`
`Fig. 9. Mass transfer corre lation for the deoxygenation of water in a
`flowing mode. Closed sy mbols. PVTMS memb rane resistance taken
`into account. Ope n sy mbols. ignoring PVTM S membrane resistance
`contribution. Cross sy mbo ls. PDMS /PPSQ co mposite membrane
`resista nce taken into acco unt.
`
`

`
`D.G. Bessarabov et al. / Journal ofMembrane Science I IJ (19961275-28-1
`
`281
`
`bber-bascd asymmetric gas-separation membranes
`re the best choice, at this stage, for such a system.
`
`,2. Hydrogen permeation
`
`Theresults of the hydrogen permeation experiments
`nthe three-channel module are presented in the Table
`i, which shows that the flux of the hydrogen permeating
`.hrough the SMV increased in the low range of the
`. quid flow rate and decreased slightly in its higher
`range.
`for hydrogen in
`The permeability coefficient (P)
`P\'T~IS is 2.0x IO- x (crrr' em/em? s cmHg) under I
`aim pressure difference [40 J. The permeability coef(cid:173)
`ncient (P) for hydrogen in water under the same con(cid:173)
`ition: can be calculated using the following equation:
`
`(7)
`
`here S is the solubility coefficient and D is the diffu(cid:173)
`Ion coefficient of hydrogen in water. The diffusion
`fficient of hydrogen in water is 5.0 X 10-" em?I s
`.J II. The solubility coefficient (5) of hydrogen in
`water at a pressure of I atrn at 200e is 2.3 X 10- 4
`em'(STP) I cm' cmHg). The calculated permeability
`)1 hydrogen in water by means ofEq. (7) is 1.2 X IO- x
`cm'(S'I'Pjcrn/crrr' s cmHg). The permeance (PII)
`t hydrogen, that is. the ratio of permeability to thick(cid:173)
`ne s. through the liquid layer in our experiments was
`12x 10-x/O.045 =2.7 X 10- 7
`(cm'(STP)/cm 2
`s
`mHg). Since the thickness of the diffusion layer in
`the PVTMS membranes was 2.3 X 10- 4 em [ 17], the
`permeance (P I I) of hydrogen in such a membrane was
`" x IO -~. The total resistance of the system to gas
`permeation ( I I QTlltal) was estimated by [37]:
`
`I
`2
`- - -+ - -
`
`QTot.J1
`
`Qf\h'mhranc
`
`QLiljuic.J
`
`(8)
`
`Ie.-
`Expenmental results for hydrogen permeation through the SMV
`omprising PVTMS membranes and operating in the flowing mode
`
`Hydrogen perrneancethrough SMV
`Icm'( STP) /cm 2 s cmHg I
`
`1.1X 10- 1>
`lA X 10- 1>
`1.4X 10- ('
`0.8 X 10- ('
`
`where I I QLiquid is the resistance of the liquid layer to
`gas permeation and 21QMcmhranc is the resistance of the
`membranes to gas permeation. It is seen from Eq. (8).
`the calculated value for the hydrogen permeance and
`experimental data in Table 5, that
`the mass transfer
`through the SMV is controlled by the diffusion in the
`liquid layer. The experimental data for the low range
`of the liquid flow rates demonstrated that the hydrogen
`permeance was larger than for stagnant liquid. but less
`than for the two polymeric membranes. The increase
`in the permeance at the lower range of the water flow
`rates can be explained by the increase in the so-called
`eddy mass diffusivity (EI» in such a system. It can be
`assumed that the presence of the turbulence promoters
`in the liquid channels increased eddy diffusivity of
`hydrogen with an increase in liquid flow rates. It can
`be assumed that the mass transfer in the SMV is char(cid:173)
`acterized by two factors acting in opposite directions:
`an increase in mass transfer with an increase in liquid
`flow rate due to increasing eddy mass diffusivity: and
`a deerease in mass transfer with an increase in liquid
`flow rate due to a convectional drift that takes place in
`the SMV operating in the flowing mode 1231.
`In our experiments we were restricted by the
`mechanical stability of the membrane three-channel
`cassette; with an increase in the liquid flow rate the
`pressure of water was also increased. When the water
`pressure, measured by manometers at the inlet of the
`three-channel system, was larger than 2.5 atm. the cas(cid:173)
`sette leaked. This could be explained by the fact that
`the total number of the channels for liquid in the new
`type of the "three-channel" cassette decreased in com(cid:173)
`parison with that number of in the "two-channel" cas(cid:173)
`sette [ 18] .
`The two-channel liquid-membrane contactor, which
`was described above. can be used not only for degas(cid:173)
`sing liquids but also for a reverse process. for example.
`oxygenation of water. Where liquid circulates between
`two contactors, gas separation of a binary mixture is
`possible. In that case, one of the gas components should
`react
`reversibly with the
`liquid phase. This was
`described earlier for the separation of ethylene l ethane
`mixture [ 18].
`
`4. Conclusions
`
`The aim of this article was to demonstrate the use of
`non-porous polymeric flat-sheet ga s-separation mcm-
`
`

`
`282
`
`D.G. Bessarabov et al. / Journal of Membrane Science //3 (/996) 275-284
`
`SP
`L
`
`spacer
`liquid
`
`Acknowledgements
`
`The large-scale gas-separation systems incorporat(cid:173)
`ing flowing-liquid membranes were manufactured in
`the course of a collaboration between the University of
`Stellenbosch, (US), (South Africa) and the Research
`of
`the OKB
`" Fine Bio-Industry",
`Division
`("TBM" ) , ( Russia, Kirishi) . The membrane system
`was manufactured partly at the US, Institute for Poly(cid:173)
`mer Science, and partly at the OKB "TBM". The
`enthusiasm and assistance of Mr. J. Blom from the
`Department of Mechanical Engineering, (US) and Mr.
`S.E . Sokolov, (OKB, "TBM") in the manufacture of
`the above-mentioned systems is greatly appreciated.
`The supplier of the PYTMS membranes was NPO
`"Plastrnass" (Russia). The supplier of the PDMS/
`PPSQ composite membranes (LESTOSIL"o) was NPO
`"Polirnersintez" (Russia).
`
`References
`
`[II M. Teramoto, H. Matsuyarna, T. Yamashiro and Y. Katayama.
`Separation of ethylene from ethane by supported liquid
`membranes contained silver nitrate as a carrier. J. Chern . Eng.
`Jpn .. 19(5 ) ( 1986) 419.
`[21 R.D. Hughe s. J.A. Mahoney and E.F . Steigelmann, Olefin
`Separation by Facilitated Transport Membranes. in N.N. Li and
`J.M. Calo (Eds. ), Recent Developments in Separation Science.
`Vol. 9. CRC Press. Cleveland. 1986. p. 173.
`131 R.W. Baker.
`I.C, Roman
`and H.K . Lonsdale. Liquid
`the production of oxygen-enriched air I.
`membranes for
`Introduction and passive liquid membranes. J. Membrane Sci..
`31 (1987) 15.
`141 B.M. Johnson. R.W. Baker. S.L. Matson. K.L. Smith . i.c
`Roman. M.E. Tuttle and H.K . Lonsdale. Liquid membranes for
`the production of oxygen-enriched air II. Facilitated-transport
`membranes. J. Membrane Sci .. 31 ( 1987) 31.
`151 H. Kreulen, e.A. Smolders. G.F. Versteeg and W.P.M. van
`Swaaij , Selective removal of H2S from sour gas with
`microporous membranes. Part II. A liquid membrane of water(cid:173)
`free tertiary amines, J. Membrane Sci .• 82 (1993) 82.
`161 Z. Qi and E.L. Cussler, Microporous hollow fibers for gas
`absorption. I. Mass transfer in the liquid. J. Membrane Sci.. 23
`(1985) 321.
`171 Z. Qi and E.L. Cussler, Microporous hollow fibers for gas
`absorption.
`II. Mass
`transfer across
`the membrane. J.
`Membrane Sci .. 23 ( 1985 ) 333.
`
`the possible
`branes in membrane-liquid contactors;
`modification of the two-channel liquid-membrane con(cid:173)
`tactor was shown. Evaluation of a novel
`large-scale
`configuration for an SMY (selective membrane valve)
`comprising non-porous gas separation membranes and
`flowing liquid membranes for the control of hydrogen
`diffusing through the SMY and water-deoxygenating
`two-channel
`experiments by means of conventional
`membrane system led to the following conclusions:
`•
`the flux could be controlled by adjusting the liquid(cid:173)
`membrane flow rate;
`the overall mass transfer coefficients were functions
`of the water flow rate; and
`the liquid-film resistance controlled mass transfer in
`both the experiments.
`
`•
`
`•
`
`5. List of symbols
`
`wetted area of the spacer (em")
`specific area of the spacer (ern - I)
`di ffusion coefficient (ern?/ s)
`eddy mass diffusivity (cm 2/s)
`mass transfer coefficient (cm/s)
`membrane length (ern)
`Henry's constant (cm3(STP) cml-lg/crrr')
`permeability (crrr'{S'TPjcm/crrr' s cmHg)
`permeance (crrr'( STP) /cm2 s cmHg)
`solubility coefficient (cm 3(STP)/cm3
`cmHg )
`volume (crrr')
`liquid velocity (cm/s)
`membrane area per volume of the
`membrane bed (em - I )
`channel height (ern)
`hydraulic diameter (ern)
`thickness of the diffusional membrane
`layer (cm)
`selectivity (-)
`porosity (- )
`kinematic viscosity (crrr'/ s)
`Sherwood number, Sh = Kh d / D (- )
`Reynolds number, Re = hdfl/ v (-)
`Schmidt number, Sc = v / D (-)
`
`A sp
`
`A v .sl'
`
`D E
`
`ll
`
`=P//
`
`K L H P Q
`
`S
`
`avn
`
`Q'
`
`E
`
`v S
`
`h
`Re
`Sc
`
`Subscripts
`OY
`Tot
`
`overall
`total
`
`

`
`D.G. Bessarabov et al. / Journal ofMembrane Science 113 (1996) 275-284
`
`283
`
`II.I.-C Yang and E.L. Cussler, Designing hollow-fiber
`contactors. AIChE J.• 32( II) (1986) 1910.
`191 Mol. Semmens. D.M. Foster and E.L. Cussler, Ammonia
`removal from water using microporous hollow fibers. J.
`Membrane Sci.. 51 ( 1990 ) 127.
`101 H. Kreulen, G .F. Versteeg. CA. Smolders and W.P.M. van
`Swaaij, Selective removal of H,S from sour gas with
`microporous membranes. Part I. Application in a gas-liquid
`system. J. Membrane Sci .. 73 (1992) 293 .
`1111 H. Kreulen, CA. Smolders. G.F . Versteeg and W.P.M. van
`Swaaij, Microporous hollow fiber membrane modules as gas(cid:173)
`liquid contact ors. Part 2. Mass transfer with chemical reaction.
`J. Membrane Sci.. 78 ( 1993) 217.
`in
`1121 S. Karoor and K.K. Sirkar, Gas Absorption Studies
`microporous hollow fiber membrane modules. Ind. Eng. Chern.
`Res.. 32 (1993) 674 .
`1131 A.B. Shelekhin and LN. Beckman. Gas separation processes in
`membrane absorber. J. Membrane set., 73 (1992) 73.
`141 D.G. Bessarabov and I.
`. Beckman. Model of mass transfer in
`the integrated liquid-membrane system for gas separation.
`Vestnik Moskovskogo Universiteta, Ser . 2. Khimia, 34(6)
`( 1993) 604 ( in Russian).
`1151 D.G. Bessarabov, D.A. Sirtsova, V.V. Teplyakov and LN.
`ethylene/ methane mixture by
`Beckman. Separation of
`membrane/absorption system with flowing liquid selective
`ab.orbent. Vestnik Moskovskogo Universiteta, Seria
`2.
`Khimia. 35(4) (1994) 385 (in Russian) .
`Jacobs and LN.
`1161 D.G. Bessarabov, R.D. Sanderson. E.P.
`Beckman. Mass transfer in absorption/membrane integrated
`hybrid systems
`for gas
`separation. Second Int. Symp..
`PMST'94. Dinkeldruk, Oldenzaal, The Netherlands. 1994.
`Abstr., pp. 59-60.
`1"1 D.G. Bessarabov, R.D. Sanderson. E.P . Jacobs. C.N. Windt
`and V.S. Gladkov. Asymmetric non-porous
`flat
`sheet
`membranes
`from polyvinyltrimethylsilane
`for
`water
`oxygenation and deoxygenating. SA J. Chern. Eng .. 6(2)
`( 1994) 26.
`1181 D.G. Bessarabov. R.D. Sanderson. E.P. Jacobs and I.N.
`Beckman. High-efficiency separation of ethylene/ethane
`mixture by a large-scale liquid-membrane contactor containing
`flat-sheet non-porous polymeric gas-separation membranes
`and a selective flowing-liquid absorbent. Ind. Eng. Chern. Res.•
`34 (1995 ) 1769.
`1191 D.G. Bessarabov , R.D. Sanderson and E.P. Jacobs. Membrane
`separation and affinity interaction . A new hybrid gas separation
`process. Proc. 7th Natl. Meet. South African lnst. Chern . Eng .•
`22-24 August. Esselen Park . South Africa. 1994. pp. 582-592.
`120 I M. Teramoto. H. Matsuyama, T. Yamashiro and S. Okamoto.
`Separation of ethylene from ethane by a (lowing liquid
`membrane using silver nitrate as a carrier. J. Membrane Sci .•
`45 (1989) 115.
`12111.N. Beckman. D.G. Bessarabov and U.R. Dzelrne, V.V.
`Teplyakov, The device for membrane gas separation. Russian
`Pat. no. 5024663/26. BOlD 61/38.1992 (in Russian).
`12211.N. Beckman. D.G. Bessarabov, V.V. Teplyakov and A.V.
`Teplyakov, Integrated membrane systems with moving liquid
`carriers for multicomponent gas separation. 3rd Int. Conf.
`
`Effective Membrane Processes. Mechanical Engineering.
`London. 1993. pp. 297-306.
`123] LN. Beckman. D.G. Bessarabov and V.V. Teplyakov, Selective
`membrane valve for ternary gas mixture separation: model of
`mass transfer and experimental test. Ind. Eng. Chern. Res.• 32
`(1993) 2017 .
`124] D.G. Bessarabov and LN. Beckman. Phenomenological theory
`of selective gas permeation in the membrane valve . Vestnik
`Moskovskogo Universiteta, Seria 2. Khimia. 34(2) (1993)
`194 (in Russian) .
`[25] LN. Beckman. Unusual membrane processes : non-steady-state
`regimes. nonhomogeneous and moving membranes. in D.R.
`Paul and Y.P. Yampolskii (Eds.). Polymeric Gas Separation
`Membranes. CRC Press. 1994. pp. 301-352 .
`(26] S. Majumdar. A.K . Guha and K.K. Sirkar, A new liquid
`membrane technique for gas separation. AIChE J.. 34(7 )
`(1988) 1135.
`(27] S. Majumdar. A. Sengupta and . K.K. Sirkar, Simultaneous
`SO,/NO separation from flue gas
`in a contained liquid
`membrane permeator, Ind. Eng . Chern . Res .• 33 ( 1994) 667 .
`128] A.K . Guha, S. Majumdar and K.K. Sirkar, Gas separation
`modes in a hollow fiber contained liquid membrane perrneator,
`Ind. Eng . Chern. Res .• 31 (1992) 593.
`[29) T. Papadopoulos and K.K. Sirkar, A modified hollow fiber
`contained liquid membrane technique for gas separation at high
`pressures. J. Membrane Sci .. 94 ( 1994) 163.
`/30] S.G. Kimura and G.F. Walmer, Fuel gas purification with
`permselective membranes. Sep . Sci . Technol ., 15 ( 1980) 1115.
`(31) A.E. Jansen. R. Klaassen. P.H.M. Feron, J.H. Hanemaaijer and
`gas
`absorption
`processes
`in
`B. Meulen, Membrane
`environmental
`applications.
`in Membrane Processes
`in
`Separation and Purification. NATO ASI Series. Series E. Vol.
`272. Kluwer, Dordrecht, 1994. pp. 343-356.
`1321 E.L. Cussler, Hollow fiber contactors, in Membrane Processes
`in Separation and Purification. NATO ASI Series. Series E.
`Vol. 272. Kluwer, Dordrecht, 1994. pp. 375-394.
`(33) T. Ahmed and MJ. Semmens. Use of sealed end hollow fibers
`for bubbleless membrane aeration: experimental studies. J.
`Membrane Sci .• 69 (1992) I.
`1341 M.S.L. Tai. I. Chua. K. u. WJ . Ng and W .K. Teo. Removal
`of dissolved oxygen in ultrapure water production using
`microporous membrane modules. J. Membrane Sci .•87 ( 1994)
`99.
`[351 T. Leiknes, MJ. Semmens and C. Gantzer, The Performance
`of a Membrane Vacuum Degasser, 3rd Int. Conf. Effective
`Membrane Processes. Mechanical Engineering. London. 1993.
`pp. 341-351.
`[361 P. Cote. J.-L. Bersilon and A. Huyard. Bubble-free aeration
`using membranes: mass transfer analysis.• J. Membrane Sci ..
`47 (1989) 91.
`[371 H. Yasuda. C.E. Lamaze. Transfer of gas to dissolved oxygen
`in water via porous and nonporous polymer membranes. J.
`Appl. Polym. ser., 16 (1972) 595 .
`138] S.R. Wickrarnasinghe, MJ . Semmens and E.L. Cussler,
`Hollow fiber modules made with hollow fiber
`fabric. J.
`Membrane Sci.. 84 ( 1993) I.
`
`

`
`284
`
`D.G. Bessarab ov et al. / Journal of Membrane Science 113 (1996) 275-284
`
`1391 G. Schock and A. Miquel, Mass transfer and pressure loss in
`spiral wound modules. Desalination. 64 (1987) 339.
`140 I V. Teplyakov and P. Meares, Correlation aspects of
`the
`selec tive gas permeabilities of polymeric materials and
`membranes. Gas. Sep . Purif., 4 ( 1990) 66 .
`
`1411 D.L. Wise and G. Houghton. The diffusi on coe fficients of ten
`slightly soluble gases in water at I0-60°C. Chern. Eng. Sci..
`21 (1966) 999.
`1421