Documen downloaded om:
This pape mus be ci ed as:
The inal publica ion is a ailable a
Copy igh
Addi ional In o ma ion
Else ie
Baumann, S.; Se a Al a o, JM.; Lobe a González, MP.; Escolás ico Rozalén, S.; Schulze-
Kueppe s, F.; Meulenbe g, WA. (2011). Ul ahigh oxygen pe mea ion lux h ough suppo ed
Ba0.5S 0.5Co0.8Fe0.2O3-del a memb anes. Jou nal o Memb ane Science. 377(1-2):198-
205. doi:10.1016/j.memsci.2011.04.050
h p://dx.doi.o g/10.1016/j.memsci.2011.04.050
h p://hdl.handle.ne /10251/76552
1/16
Ul ahigh Oxygen Pe mea ion Flux h ough Suppo ed
Ba0.5S 0.5Co0.8Fe0.2O3- Memb anes
Jou nal o Memb ane Science (2011) (accep ed 24/04/2011)
doi: 10.1016/j.memsci.2011.04.050
S. Baumann1, J. M. Se a2,*, M. P. Lobe a
2, S. Escolás ico2, F. Schulze-Küppe s1, W. A.
Meulenbe g1
1Fo schungszen um Jülich GmbH, Ins i u e o Ene gy and Clima e Resea ch IEK-1, Leo-B and -S . 1, D-52425 Jülich, Ge many
2Ins i u o de Tecnología Química (Uni e sidad Poli écnica de Valencia – Consejo Supe io de In es igaciones Cien í icas), A . Na anjos s/n,
E-46022 Valencia, Spain.
ABSTRACT
Oxygen anspo memb anes made o Ba0.5S 0.5Co0.8Fe0.2O3- (BSCF) we e manu ac u ed by
ape cas ing and co- i ing. The disk-shaped memb anes consis ed o a op gas igh laye
(70 µm hick) and a po ous subs a e (830 µm hick) wi h 34% open po osi y. The a ia ion
o he pe mea ion ope a ion condi ions allowed (i) he iden i ica ion o he di e en
limi a ions s eps in he pe mea ion p ocess, i.e. bulk oxygen ion di usion, ca aly ic su ace
exchange and gas phase di usion in he memb ane compa men s and po ous subs a e; and
(ii) he ul ima e op imiza ion o he oxygen lux. The a iables conside ed in he sys ema ic
pe mea ion s udy included he inle gas low a e o he sweep and ai eed, he empe a u e
and he na u e o he oxygen eed gas (ai o pu e oxygen). Mo eo e , he in luence o he
deposi ion o a ca aly ic ac i a ion laye (17 µm hick) made o BSCF on op o he hin
gas igh laye was in es iga ed. As a esul o his pa ame ic s udy, unp eceded oxygen lux
alues we e achie ed, i.e. a maximum lux o 67.7 ml(STP) min-1cm-2 was ob ained a
1000 ºC using pu e oxygen as he eed and a gon as he sweep, while a lux o 12.2 ml(STP)
min-1cm-2 a 1000 ºC was ob ained when ai was used as he eed.
Keywo ds: oxygen anspo memb ane; suppo ed memb ane; oxygen lux; BSCF; oxygen
sepa a ion
* Co esponding au ho
Fax: +34 963 877809; Tel: +34 963 879448; E-mail: jmse a@i q.up .es
2/16
1. In oduc ion
One impo an s a egy o educing CO2 emissions, while he ene gy demand is
inc easing, is ca bon cap u e and s o age (CCS). Di e en ossil uel powe plan concep s o
CCS a e cu en ly being de eloped. Examples include pos -combus ion, p e-combus ion and
oxy uel. In oxy uel powe plan s, he ossil uel is combus ed using pu e oxygen o a
ni ogen- ee gas s eam en iched wi h oxygen. The esul o his concep is an o -gas
con aining p ima y CO2 and H2O, i.e. 90-95% CO2 in he d ied lue gas. The CO2 can hen be
cap u ed mo e easily han when ai is used in he combus ion p ocess, which leads o 10-14%
CO2 in he d ied lue gas. [1] The oxygen equi ed o his concep can be p o ided by
di e en me hods, o which ce amic oxygen anspo memb anes* (OTMs) ha e he lowes
e iciency losses [2], especially when eci cula ed lue gas is used as sweep gas.
OTMs consis o gas igh mixed ionic-elec onic conduc o s (MIEC) and allow oxygen
di usion ia oxygen acancies in he c ys al la ice. The e o e, he selec i i y o he
memb anes is in ini e, which means ha hey p o ide pu e oxygen. The mos p omising
ma e ials o OTMs a e pe o ski es wi h he o mula ABO3- [3]. In las decade, in ensi e
esea ch has been dedica ed o he p epa a ion and cha ac e iza ion o MIEC memb anes [4-
12]. The ma e ials showing he highes oxygen pe meabili y a e based on
Ba0,5S 0,5Co0,8Fe0,2O3- (BSCF), which was i s epo ed by Shao e al. [13] and has a ac ed
g ea in e es ecen ly [14-18]. The he modynamic d i ing o ce o oxygen anspo h ough
a MIEC memb ane is he oxygen chemical po en ial g adien along he hickness o he
memb ane, which is gi en by he ope a ing condi ions. The oxygen pe mea ion lux based on
bulk di usion can be desc ibed by he Wagne Equa ion [3]
22
2
'
2
''
2
2ln)(
16 pOdpO
LF
RT
JO
O
P
P
ambO 


(1)
whe e JO2 is he oxygen pe mea ion lux in mol.m-2.s-1, R is he gas cons an , F is he Fa aday
cons an , L is he memb ane hickness, amb

is he ambipola conduc i i y, and pO2’ and pO2”
a e he oxygen pa ial p essu es a he high p essu e side and low p essu e side, espec i ely.
In o de o inc ease he oxygen pe mea ion lux h ough a memb ane, he hickness
should be as low as possible. When he hickness becomes e y low, a po ous suppo is
needed o mechanical s abili y, pa icula ly in he case o plana memb anes. This has led o
ecen inc eased in e es in he de elopmen o suppo ed memb anes. The memb anes ha
ha e been in es iga ed a y widely in e ms o memb ane hickness (10 µm o 500 µm) and
ma e ial [19-26] ( able 1). The pe mea ion enhancemen s ha ha e been epo ed a e no mally
much less han he alues p edic ed by he Wagne equa ion (1), assuming di ec ecip ocal
beha iou o memb ane hickness and pe mea ion a e. This is due o he ac ha Wagne ’s
* Also called ionic anspo memb anes (ITMs)
3/16
heo y is only alid o solid s a e di usion; i does no include su ace exchange kine ics (O2
adso p ion, dissocia ion and educ ion) o o he gas-phase phenomena, which a leas pa ly
domina e pe mea ion h ough hin memb anes. Fu he mo e, a limi ing in luence o he
po ous suppo has o be conside ed [26].
This wo k p esen s he p epa a ion o asymme ic memb anes made o BSCF. The
sys ema ic s udy o he memb ane ope a ion a iables (inle gas low a e a he wo
memb ane compa men s, empe a u e and he na u e o he oxygen eed gas) and hei
in luence on he oxygen lux is discussed. Mo eo e , he in luence o he applica ion o an
ac i a ion laye on op o he hin gas igh laye is in es iga ed. The oxygen lux is op imized
by elimina ing he di e en p ocess limi a ions s ep by s ep.
2. Expe imen al
2.1. Memb ane p epa a ion
Bo h he memb ane laye and he suppo we e manu ac u ed by ape cas ing using
BSCF. This app oach has se e al ad an ages, such as pe ec chemical compa ibili y and he
same he mal expansion o he wo laye s [22]. The comme cial BSCF powde used
(T eibache Indus ie AG, Aus ia) exhibi ed an a e age g ain size o 1.7 µm. In he suppo ,
co n s a ch (Ca gill, Ge many) was used as he po e o me wi h a pa icle size in he ange o
2-30 µm [27].
The slu y p epa a ion p ocedu e is shown in igu e 1 and was pe o med acco ding o
[28]. Two slu ies we e p epa ed: he slu y o he suppo laye con ained 20% co n s a ch in
ela ion o he o al solid con en , whe eas he slu y o he memb ane laye did no con ain
any po e o me . Sequen ial ape cas ing was used o p epa e he memb ane. Fi s , he
memb ane laye was cas in wo s eps wi h a d ying s ep in be ween using cas ing gaps o
50 µm and 100 µm, espec i ely. A e d ying, he suppo laye was cas on op o his laye
wi h a cas ing gap o 1.9 mm. This laye was hen d ied and he samples we e cu ou o he
g een ape, debinde ed and sin e ed in ai a 1100°C o 3 h.
The oxygen ac i a ion laye s we e p epa ed ia sc een p in ing. The sc een p in ing
ink consis ed o BSCF powde p o ided by F aunho e IKTS (He msdo , Ge many). I was
ball-milled in an ace one suspension p io o p epa a ion o he sc een-p in ing inks and an
a e age pa icle size o 250 nm was achie ed. A ypical addi i e con aining 94 w .% e pineol
and 6 w .% e hylene cellulose was used. Fu he homogeniza ion was conduc ed using a
h ee- oll mill. In o de o comple e homogeniza ion, his s ep was epea ed h ee imes.
Mo eo e , g aphi e (Ald ich) was used as he po e o me in he sc een-p in ing ink. The
p in ed laye was sin e ed a 1010 ºC.
4/16
The mic os uc u e was in es iga ed using scanning elec on mic oscopy (SEM) (Zeiss
Ul a 55), and elemen al analysis was ca ied ou wi h ene gy-dispe si e X- ay spec oscopy
(EDS) (INCA, Ox o d). The po osi y o he memb anes ob ained was in es iga ed by ligh
mic oscopy and quan i a i e image analysis using comme cial so wa e, i.e. ImageJ and
analySIS. Gas igh ness was measu ed using He leakage (P ei e acuum).
2.2. Oxygen lux measu emen s
Oxygen pe mea ion s udies we e ca ied ou in a lab-scale qua z eac o . Syn-
he ic ai (21% / O2 in he eed s eam) o pu e oxygen was ed in o he oxygen- ich cham-
be , while a gon was used as he sweep gas on he pe mea e side. Bo h gases we e ed a a -
mosphe ic p essu e. Inle gases we e p ehea ed in o de o ensu e he co ec gas empe a u e
o con ac wi h he memb ane su ace. This is pa icula ly impo an when high gas low
a es a e employed. All s eams we e indi idually mass low con olled. The empe a u e was
measu ed by a he mocouple a ached o he memb ane. A PID con olle main ained empe a-
u e a ia ions wi hin 2 ºC o he se poin . The samples consis ed o gas igh suppo ed BSCF
memb anes and memb ane gas leak- ee condi ions we e achie ed using gold ings on bo h
sides o he memb ane, which we e hea ed o 1010 °C o 4 hou s immedia ely p io o he
measu emen . The pe mea e was analysed a s eady s a e by online gas ch oma og aphy using
a mic o-GC Va ian CP-4900 equipped wi h Molsie e5A, Po a-Plo -Q glass capilla y, and CP-
Sil modules. Memb ane gas leak- ee condi ions we e ensu ed by con inuously moni o ing he
ni ogen concen a ion in he p oduc gas s eam (jus be o e and a e swi ching o a pu e O2
eed). An accep able sealing was achie ed when he a io be ween he oxygen low leak and
he oxygen lux was lowe han 1%. The da a epo ed he e we e achie ed a s eady s a e a e
one hou in he eac ion s eam. Each es was epea ed h ee imes o minimize he analysis
e o . The expe imen al analy ical e o was below 0.5 %.
Oxygen pe mea ion was de e mined in he empe a u e ange o 700-1000 ºC. The
esul s ob ained a di e en ope a ing condi ions a e shown below. JO2 was s udied o a ious
sweep gas low a es (i.e. di e en oxygen pa ial p essu e on he pe mea e side, pO2”) and
di e en pO2’ (syn he ic ai o pu e oxygen). Oxygen pe mea ion measu emen s we e ca ied
ou wi h he memb ane laye and he suppo on he sweep and he eed side, espec i ely.
This si ua ion was chosen because concen a ion pola iza ion limi a ions in he suppo can
easily be o e come by using pu e oxygen as eed. In con as , i would be impossible o easily
elimina e such e ec s a he sweep side. The e o e, he po en ial o he hin suppo ed
memb ane laye and he su ace ac i a ion laye can be assessed ollowing his app oach.

5/16
3. Resul s and Discussion
3.1. Memb ane assembly mic os uc u e
A p elimina y sin e ing s udy o bo h apes wi h di e en ypes and amoun s o po e
o me s (suppo ) and apes wi hou po e o me (memb ane) was pe o med aiming a a
po ous suppo and a gas igh memb ane laye , espec i ely, a he same sin e ing
empe a u e. The sin e ing sh inkage o he wo apes has been adap ed in such a way ha a
maximum ma ch exis ed, leading o low bending o he co-sin e ed sample acked by lase
opog aphy. The bes esul s we e ob ained when using 20 w % co n s a ch as a po e o me
and subsequen ly co- i ing a 1100°C o 3 hou s in ai . Fo he disk-shaped samples (20 mm
in diame e ), gas igh ness was con i med by measu ing he He leak, e ealing a alue o 6.10-
6 mba l cm-2 s-1.
Figu e 2 p esen s he SEM analysis o ac u e c oss-sec ions o he samples a e
pe mea ion es s. Figu e 2a shows an o e iew o he co-sin e ed memb ane assembly. The
po osi ies o he memb ane laye and he suppo laye we e 3% and 34%, espec i ely, as
de e mined by image analysis. The po e size o he subs a es anged om 5 o 25 µm while
he closed po es o he dense memb ane laye had an a e age po e size o app ox. 3 µm (see
Figu e 2b). The BSCF g ain size a e sin e ing a 1100ºC was la ge, ypically abo e 10 µm.
A de ailed analysis a high magni ica ions sugges ed ha mos o he po es in he op dense
laye we e occluded in he g ain in e io and no p e e en ially loca ed a g ain bounda ies.
The hickness o he gas igh memb ane laye was 70 µm and i was e y homogenous. The
whole memb ane assembly was 900 µm hick. The subs a e was composed o e y well-
bound and sin e ed BSCF g ains.
The o ma ion o a ine-g ain po ous laye wi h a hickness below 100 nm was obse ed on
op o he dense BSCF laye . I was no ed ha his po ous laye was no well a ached o he
dense memb ane su ace. This po ous laye was no p esen be o e he es ing and o igina ed
du ing ope a ion, which in ol ed con inuous sweeping wi h a gon (al hough he a e age
oxygen pa ial p essu e was ypically abo e 0.001 due o he high oxygen lux). These su ace
deposi s a e mos likely caused by ouling, i.e. he deposi ion o pa icles (SiO2, C Ox, FeOx,
e c.) om he piping, he qua z eac o o he sealing ma e ial. The su ace seg ega ion o
seconda y phases om he BSCF bulk [29] may also ha e aken place. In gene al, he
memb anes do no show any appa en deg ada ion o damage.
Figu es 2c and 2d show he memb ane wi h an ac i a ion laye composed o po ous
BSCF. The oxygen ac i a ion laye had an open mic os uc u e and a homogeneous hickness
o 17 µm. Figu e 2d and Figu e 3 show highe magni ica ion images o he po ous laye . The
pa icle size was well below 1 µm and he p esence o big plana oids (mac opo es) was
obse ed. These po es we e p oduced by he inco po a ion o g aphi e pla ele s in he p in ing
6/16
ink and he subsequen emo al du ing sin e ing. The in oduc ion o such mac opo es aimed
o enhance gas exchange h ough he laye . Fu he mo e, he po ous laye was ound o be
s able du ing he whole high- empe a u e oxygen pe mea ion measu emen .
3.2. Oxygen pe mea ion
3.2.1. T anspo mechanisms model
Oxygen pe mea ion comp ises a e y complex se o di e en anspo mechanisms,
which can be in e p e ed as a se ies o esis ances (Figu e 4). Depending on he ope a ing
condi ions, he d i ing o ce can be dec eased due o a deple ion o accumula ion o O2 in
po ous laye s, i.e. suppo and ac i a ion laye s, o e en in he gas phases. This gi es ise o a
lowe oxygen concen a ion g adien in he bounda y laye . This phenomenon is e e ed o as
concen a ion pola iza ion and i has se ious de imen al e ec s in he memb ane sepa a ion
p ocess. In he eed gas, concen a ion pola iza ion can occu especially a high pe mea ion
a es due o oxygen deple ion in he gas phase abo e he suppo (RCP1). This e ec is
expec ed o be e en s onge wi hin he po es o he suppo , whe e molecula di usion is he
p edominan anspo mechanism (RSUPPORT). A he memb ane su ace, he oxygen has o
adso b, educe, and dissocia e in o de o o m oxide ions, which ha e o inco po a e in o he
pe o ski e la ice. This mul is ep p ocess is summa ized in one esis ance RS1. The oxygen
ions di use h ough he pe o ski e bulk and g ain bounda y (RSOLID). Then, he mul is ep
su ace exchange (RS2) occu s and he oxygen is deso bed. Gene ally, he gas-phase esis ance
a he pe mea e side may be negligible compa ed o ha o solid s a e di usion and exchange
eac ions, bu when he oxygen pe mea ion lux is su icien ly high, concen a ion
pola iza ion in he gas phase (RCP2) becomes signi ican . All o hese esis ances a e
dependen on he espec i e local oxygen pa ial p essu e, which a e no di ec ly de ec able.
The empe a u e dependence o he oxygen pe mea ion lux h ough a ba e, suppo ed
BSCF memb ane was s udied (Figu e 5) using wo di e en oxygen concen a ions in he eed
s eam, i.e. syn he ic ai and pu e oxygen. The esul s indica ed A henius beha iou al hough
he appa en ac i a ion ene gies (Ea,a) o oxygen anspo changed as a unc ion o eed gas
na u e and/o empe a u e, which sugges s ha he a e-limi ing s ep changes acco ding o he
empe a u e ange in es iga ed. This beha iou is cha ac e is ic o pe o ski e- ype oxides,
such as BSCF. In he high- empe a u e ange, oxygen pe mea ion is ypically limi ed by oxy-
gen ions di using h ough he pe o ski e bulk and g ain bounda y, while a lowe empe a-
u es i is mo e limi ed by su ace s eps. Mo eo e he Ea,a o oxygen su ace exchange is
highe han ha o he oxygen ion di usion [3,9,13-15,18]. Howe e , his simpli ied pic u e
does no p ope ly desc ibe he pe mea ion p ocess in suppo ed memb anes (e.g. pe mea ion
esul s in Figu e 5) when high oxygen luxes a e ob ained. In he p esen case, all o he men-
ioned esis ances may con ibu e o he o e all p ocess esis ance.
3.2.2. In luence o ca aly ic laye
7/16
The memb ane su ace was modi ied by deposi ing a po ous ac i a ion laye o BSCF
aiming o imp o e he oxygen lux pe mea ed (JO2). The plo s in Figu e 6 show he oxygen
pe mea ion luxes h ough he ba e memb ane and he coa ed memb ane in he empe a u e
ange o 700-1000 ºC o di e en sweep gas low a es when syn he ic ai was used as he
eed gas. As expec ed, he esul s indica ed A henius beha iou as a unc ion o he
empe a u e. Speci ically, hey showed wo anges o appa en ac i a ion ene gy (Ea,a) o
oxygen pe mea ion (an es ima ion o he Ea,a is shown in able 2). The c i ical empe a u e o
he a e-limi ing s ep change was a ound 800 ºC. This change was caused by he empe a u e
dependence o he cha ac e is ic hickness Lc below which he su ace exchange kine ics
become mo e and mo e a e limi ing [3]. The su ace exchange kine ics possessed a highe
ac i a ion ene gy han solid s a e di usion, esul ing in a highe appa en ac i a ion ene gy in
he low- empe a u e egion. Fu he mo e, he su ace exchange coe icien k inc eased wi h
inc easing pO2 [3]. The e o e, ca aly ic ac i a ion was expec ed o be mo e e ec i e a he
pe mea e side. A he eed side, pa o he po ous suppo ac ed as an ac i a ion laye . The
hickness o his ac i e pa close o he memb ane laye was unce ain. The su ace ac i a ion
pe o mance o he suppo s uc u e a ached o he dense memb ane laye depends on
whe he he oxygen can pe mea e h ough he solid (pe o ski e) and he gas phase o he
suppo . Thus, i changes inhe en ly wi h he measu ing condi ions, e.g. T, pO2. In all cases,
he p esence o a ca aly ic laye on he memb ane pe mea e side imp o ed he JO2 eached and
educed he Ea,a, especially in he low- empe a u e ange. Se e al s udies ha e shown ha
su ace modi ica ion leads o an imp o ed su ace exchange a e and a co esponding inc ease
in he JO2. da Cos a e al. inc eased he JO2 ob ained wi h a BSCF hollow ib e a e su ace
modi ica ion using noble me als [11, 30]. The applica ion o a po ous laye made o a
compa ible ma e ial can also be conside ed as a way o inc easing he a ea a ailable o he
su ace exchange eac ions [31-35].
3.2.3. In luence o gas low a es
Oxygen pe mea ion lux is s ongly in luenced by he inc ease in gas low a e. This is
a ibu ed o (i) he educ ion in he concen a ion pola iza ion esis ance a bo h memb ane
sides; and (ii) he inc ease in he d i ing o ce, due o ei he he highe dilu ion o he
pe mea ed oxygen o he highe pO2 in he deple ed ai s eam.
When he gas space eloci y is inc eased, he luidodynamic beha iou o he
expe imen al se -up imp o es, and he e ec o he gas pola iza ion diminishes. Figu e 7
shows he oxygen pe mea ion lux eached a 900 ºC a a ious sweep gas low a es (QSweep)
o wo ai low a es (QFeed). In all cases, a ise in he ai low a e led o a signi ican inc ease
in he JO2. This can be asc ibed o a lowe gas concen a ion pola iza ion e ec a he ai side
(RCP1), and be e molecula di usion in he po es o he suppo (lowe RSUPPORT) o an
inc eased ai low a e. Mo eo e , QSweep a ia ions in luenced he oxygen pa ial p essu e a
he pe mea e side ( ''
2
O
P). An inc ease in QSweep educed ''
2
O
P (i.e. he o e all d i ing o ce in he
pe mea ion p ocess has inc eased) and imp o ed he luid dynamics. As a esul , he oxygen
pe mea ion lux imp o ed subs an ially wi h inc easing sweep gas low a es. Mo eo e ,
8/16
when RCP2 was subs an ially educed by inc easing QSweep, one o he majo emaining
esis ances was RS2, which became limi ing. Consequen ly, he e ec o he deposi ion o a
ca aly ic laye becomes mo e impo an wi h inc easing sweep gas low a es, as can be seen
in Figu e 6. In addi ion, Figu e 8 ( op) shows he oxygen lux ob ained wi h he su ace-
ac i a ed memb ane as a unc ion o QSweep and empe a u e. I can be asce ained ha he
inc ease in he sweep low a e is e y bene icial o he oxygen pe mea ion, especially a he
highes empe a u es when he highes oxygen luxes a e eached.
3.2.4. In luence o po ous suppo
Since molecula di usion is he main anspo mechanism in he suppo po es, he
s uc u e o he po ous suppo is c ucial o he oxygen pe meabili y in he suppo ed hin-
ilm memb anes. The e o e, i is impo an o e alua e he in luence o he suppo on he
oxygen pe mea ion p ocess. Suppo e ec s due o he gas anspo in he po e sys em can be
excluded when pu e oxygen is used ins ead o ai as he eed gas (Figu e 8 (bo om) and
Figu e 9). A emendous inc ease in pe mea ion lux was obse ed when he eed gas was
swi ched om ai o oxygen, al hough he o e all oxygen pa ial p essu e g adien ln
(pO2’/pO2”) dec eased sligh ly om 2.05 o ai o 1.99 o pu e oxygen. The use o pu e
oxygen as he eed gas allowed p e en ing limi a ions asc ibed o gas anspo h ough he
suppo (p incipally RSUPPORT) and he e o e he esul s o Figu e 9 only e lec he in luence
o he sweep gas low a e. Figu e 8 shows he s onge e ec caused by a ying he QSweep a
high empe a u es when pu e oxygen is ed in o he sys em (no e he log scale), i.e. he
inc ease in JO2 is much la ge wi h inc easing QSweep. In his case, he gas concen a ion
pola iza ion e ec s on he pe mea e side (RCP2) may be e y se e e due o he e y high
oxygen lux achie ed unde hese condi ions. Howe e , a lowe empe a u es, he
imp o emen was no as good, and i appea ed ha su ace exchange (RS2) limi s he p ocess
unde hese condi ions.
3.2.5. Op imiza ion o oxygen lux
Taking in o accoun he di e en anspo mechanisms in oxygen pe mea ion
h ough suppo ed BSCF memb anes, he bes ope a ing condi ions we e selec ed. The
oxygen pe mea ion measu emen s we e ca ied ou using ai o pu e oxygen as he eed gas.
The la e p o ides an oxygen pa ial p essu e o 1 a m and, hus, he oppo uni y o exclude
he esis ances RCP1 and RSUPPORT. Fu he mo e, RS1 is minimized since he su ace exchange
a e ( n
O
pk 2
) is maximized. On op o he dense memb ane laye , a hin po ous oxygen
ac i a ion laye was applied. This laye consis ed o he memb ane ma e ial BSCF and
enabled a dec ease in he su ace exchange esis ance (RS2). Mo eo e , RCP2 was negligible
due o he use o high low a es o he sweep gas. Consequen ly, he pe mea ion a e
inc eased signi ican ly when he eed gas was changed om ai o oxygen. The g ea e
enhancemen in he case o oxygen is consis en wi h he model based on consecu i e
esis ances desc ibed abo e. As he esis ances RCP1 and RSUPPORT we e negligible and RS1
15/16

16/16
Table 2. Es ima ion o he appa en ac i a ion ene gy (Ea,a) (kJ mol-1). QAi = 300 ml(STP)
min-1.
Ba e O2 ac i a ion laye
QA [ml(STP) min-1] Low T High T Low T High T
200 51.7 32.0 50.9 27.6
300 69.2 35.9 47.6 27.6
400 86.4 36.4 48.1 28.8
1/11
FIGURE CAPTIONS
Fig. 1. Slu y p epa a ion o ape cas ing.
Fig. 2. F ac u e c oss-sec ions (SEM pic u es) o wo memb anes a e oxygen pe mea ion
measu emen s: (a-b) ba e and (c-d) wi h an oxygen ac i a ion laye .
Fig. 3. De ails o he mo phology o he oxygen ac i a ion laye a e he pe mea ion es .
F ac u e c oss-sec ion (SEM pic u e) o he memb ane a highe magni ica ion.
Fig. 4. Concen a ion p o iles ac oss he asymme ic memb ane hickness and he
co esponding model esis ances.
Fig. 5. Tempe a u e dependence o he oxygen pe mea ion lux h ough he ba e, suppo ed
BSCF hin- ilm memb ane. Syn he ic ai (21 % / O2) o pu e oxygen in he eed s eam.
QFeed= 300 ml(STP) min-1, QSweep = 300 ml(STP) min-1.
Fig. 6. Tempe a u e dependence o he oxygen pe mea ion lux h ough he suppo ed BSCF
hin- ilm memb anes. Syn he ic ai (21 % / O2) in he eed s eam. QFeed= 300 ml(STP) min-
1. The educ ion o pola iza ion e ec s as a esul o he QSweep inc ease makes i possible o
su ace exchange o become limi ing and, he e o e, he use o an ac i a ion laye has a much
s onge impac on he inal oxygen lux.
Fig. 7. E ec o he ai low a e. BSCF hin- ilm memb ane wi h an O2 ac i a ion laye .
Di e en sweep gas low a e. T = 900 ºC; syn he ic ai (21 % / O2) in he eed s eam.
Fig. 8. Tempe a u e dependence o he oxygen pe mea ion lux h ough he BSCF hin- ilm
memb ane wi h an O2 ac i a ion laye . Di e en sweep gas low a e. QFeed = 300 ml(STP)
min-1.
Fig. 9. E ec o he po ous suppo . BSCF suppo ed hin- ilm memb ane wi h an O2
ac i a ion laye . T=900 ºC; QFeed = 300 ml(STP) min-1.
Fig. 10. Tempe a u e dependence o he oxygen pe mea ion using oxygen as a eed and
400 ml(STP) min-1 A as a sweep gas.
2/11
Figu e 1
Mixing
Sol en s
Mixing
Ba
0,5
S
0,5
Co
0,8
Fe
0,2
O
3
Powde Po e o me
(S a ch)
Mixing
Plas icize s Binde
Si ing
De-ai ing
Cas ing
Addi i es
3/11
Figu e 2
4/11
Figu e 3

5/11
Figu e 4
pO2’
pO2’’
Pe mea e side (2)Feed side (1)
R
CP1
R
SUPPORT
R
S1
R
SOLID
R
S2
R
CP2
S1 S2
6/11
Figu e 5
0.75 0.80 0.85 0.90 0.95 1.00 1.05
11
2
4
6
8
1010
20
40
60
pO2=0.21
pO2=1
1000/T (K-1)
JO2 (ml(STP) min-1 cm-2)
Ba e
7/11
Figu e 6
0.75 0.80 0.85 0.90 0.95 1.00 1.05
2
4
6
8
10
JO2 (ml(STP) min-1 cm-2)
1000/T (K-1)
Ba e
Ac i a ion Laye
QSweep = 200 ml(STP) min-1
0.75 0.80 0.85 0.90 0.95 1.00 1.05
1
2
4
6
8
10
20
JO2 (ml(STP) min-1 cm-2)
QSweep = 300 ml(STP) min-1
1000/T (K-1)
Ba e
Ac i a ion Laye
0.75 0.80 0.85 0.90 0.95 1.00 1.05
1
2
4
6
8
10
20
JO2 (ml(STP) min-1 cm-2)
1000/T (K-1)
QSweep = 400 ml(STP) min-1
Ba e
Ac i a ion Laye
Ai eed
Ai eed
Ai eed
8/11
Figu e 7
200 300 400
2
4
6
8
10
12
JO2 (ml(STP) min-1 cm-2)
QSweep (ml(STP) min-1)
QFeed=200 ml(STP) min-1
QFeed=300 ml(STP) min-1
wi h ac i a ion laye and ai eed