1
FC-10-1019 An epa a
In luence o Di e en Aspec s o he SOFC Anode En i onmen on
he Oxida ion Beha iou o Po ous Samples made o C o e
I. An epa aa*, M. Ri asa, I. Villa eala, N. Bu gosb, F. Cas ob
a) Ike lan- Ene gía, Pa que Tecnológico de Ála a, Juan de La Cie a,1 E-
01510 Miñano, Ála a, Spain
b) CEIT Ma e ials Depa men , Pº Manuel La dizabal, 15 E-20018 San
Sebas ian Spain
*co esponding au ho :
I. An epa a*, Ike lan- Ene gía, Pa que Tecnológico de Ála a, 01510 Miñano, Spain
el. +34 945 297032, ax: +34 945 296926, e-mail: [email p o ec ed]
ABSTRACT
C o e can be conside ed as he e e ence in e connec ma e ial in SOFCs wo king unde
800ºC. Thanks o i s he mal expansion coe icien (TEC), i is sui able o eplace ce amic
componen s, such as he in e connec and he me al suppo , and i can be cos e ec i e.
Se e al esea ch g oups, including Ike lan, ha e used po ous subs a es wi h he same
composi ion as C o e (PM om H.C. S a ck GmbH) as he me al suppo o hei SOFC
This is he accep ed manusc ip o he a icle ha appea ed in inal o m in Jou nal o Fuel Cell Science
and Technology 7(6) : (2010) // A icle ID 061010 , which has been published in inal o m a h ps://
doi.o g/10.1115/1.4001764. © 2010 Ame ican Socie y o Mechanical Enginee s unde CC-BY dis ibu ion
license (h p://c ea i ecommons.o g/licenses/by/4.0/)
2
FC-10-1019 An epa a
cells. The aim o his s udy is o de e mine he e ec o ce ain a iables ( ime,
empe a u e, apou con en , cycling, po osi y and cu en low), while o he aspec s a e
cons an (sample composi ion and pa icle size and shape).
KEYWORDS: SOFC, me al suppo , oxida ion, po osi y
ABBREVIATIONS: TEC, he mal expansion coe icien ; PM, powde me allu gy; FU, uel
u iliza ion; APU, auxilia y powe uni .
1 INTRODUCTION
The app oach o Ike lan’s SOFC p ojec [1] is simila o hose o o he esea ch uni s [2],
al hough i has planned a ubula ins ead o a plana geome y o he p o o ype and a
di e en inal applica ion in domes ic gene a o s ha p oduce HSW and elec ici y. The
basis is an ope a ing empe a u e o below 800ºC, so ha ce ain expensi e ce amic
componen s, such as he suppo , can be eplaced wi h me al componen s.
The e a e a many e e ences on dense me allic ma e ials in ca hode o anode oxidizing
a mosphe es, bu e y li le on he oxida ion o po ous ma e ials [3], and e en less on
po ous ma e ials in SOFC anode condi ions. The ma e ials used in some o hese s udies
we e C o e (Thyssenk upp VDM GmbH) [2, 4], and ITM (Plansee) [5]. F om he i s s udy,
i was concluded ha o a li e ime o 5000h a 800ºC he su ace a ea o he po ous
C o e mus be 0.04 m2/g o less [2].
A Ike lan, p elimina y es s we e ca ied ou in ai o dense and po ous samples o
Fe70/C 30 [3]. One o he conclusions om his s udy was ha po osi y is one o he main
issues o po ous samples in oxidizing a mosphe es. The e is cu en ly no uni e sally
accep ed uel cell indus y s anda d on how po osi y should be measu ed. A chimedes
3
FC-10-1019 An epa a
measu emen s ypically esul in sligh ly lowe po osi y alues. Me cu y po osime y esul s
in signi ican ly lowe alues [6]. Po osi y calcula ion h ough geome y and weigh was
used, al hough also me cu y po osime y da a was a ailable.
The aim o he i s pa o he p esen wo k is o de e mine he e ec on he beha iou o
po ous subs a es in anode oxidizing condi ions o ce ain a iables ( ime, empe a u e,
apou con en , cycling), while o he aspec s a e cons an (sample composi ion, pa icle
size, po osi y o no cu en lowing).
Fo sa e y easons, many esea ch g oups use H2 dilu ed in A as he oxida ion
a mosphe e, because he pa ial p essu e o O2 (pO2) is he same [7] and me al beha iou
agains oxida ion is usually a he modynamic unc ion o pO2 [8]. Addi ional es s in H2
dilu ed in A we e pe o med o compa e wi h he esul s om he i s pa wi h pu e H2
mixed wi h H2O.
Fuel gas supply o he anode h ough he me al suppo mus emain adequa e du ing any
oxida ion ha occu s du ing ope a ion [6] and su ace a ea is a key a iable in he
oxida ion p ocess o a po ous sample [2]. Thus, he pe meabili y and su ace a ea o all
he samples we e cha ac e ized using di e en echniques, and da a was compa ed in
o de o de ine he mos sui able.
In he second pa , po osi y and cu en low we e aken in o accoun as new a iables.
One sample wi h a lowe po osi y was es ed.
Like e e y elec ochemical eac ion, co osion consis s o he in e change o elec ons.
Elec ic cu en has minimal e ec on he composi ion o he oxide scales. Howe e ,
because he me al ion di usion a e will be accele a ed by he elec ical ield, he g ow h
a e o oxide scales on he nega i e side we e highe han hose on he posi i e side [9].
Thus, he cu en in a uel cell ends o slow down he g ow h a e o he oxide on he
4
FC-10-1019 An epa a
ca hode side and accele a e i on he anode side, bu he e ec s a e small o an elec onic
conduc o [10].
2 EXPERIMENTAL
Po ous subs a es we e consolida ed a high empe a u e in hyd ogen un il enough
mechanical p ope ies wi hou b eaking he samples we e ob ained (70% po osi y). C o e
powde (H.C. S a ck GmbH) was used as a aw ma e ial; no o he sample composi ion
was es ed. In addi ion, he ollowing aspec s we e cons an ; pa icle shape (i egula ) and
pa icle size. Long ubes we e consolida ed and samples o 1cm (H) x 1,4cm (Øex ) x 1cm
(Øin ) we e cu o he oxida ion s udy. The same ype o samples we e used o iso he mal
and cyclic expe imen s and o H2 dilu ed in A .
Oxida ion es s we e de ined ollowing a s anda d p ocedu e o SOFC in e connec [11].
A e measu ing and weighing (GR-200 om A&D Ins umen s LTD), he samples we e
in oduced in o he o en (o en and con olle om Ca boli e). When ai was used as he
oxidizing a mosphe e, no special equipmen was needed. Bu when using H2-H2O as he
oxidizing a mosphe e, he samples we e in oduced in o he o en inside a op-end alumina
ube (4 cm inne diam., Coo sTek, Inc.). A low o 100ml/min o pu e H2 (H2 low eloci y
o 4 mm/s a 800ºC, 10 imes he ecommended maximum [11], e en highe i humidi y is
aken in o accoun ) was used, while in he case o A -10%H2, 25ml/min was used.
Pu e H2 was o ced o bubble in o he humidi ica ion sys em (HS-AWF om
FuelCellS o e.com). This sys em was ed wi h wa e om a pu i ie (AP-3 model om
Pu agua Sys ems). The low a he ou le o he humidi ica ion sys em was supposed o be
sa u a ed, so, by con olling he empe a u e o he wa e , i was possible o con ol he
H2/H2O a io. The inal humidi y a he ou le o he humidi ica ion sys em was no
5
FC-10-1019 An epa a
measu ed. The whole pipe om his poin o he alumina ube inle was hea ed o e 100ºC
o a oid condensa ion.
The equipmen o me cu y po osime y (Au oPo e IV om Mic ome i ics and Au oPo e IV
9500 so wa e) was wo king om 50 o 60000psi, so ha po es be ween 0.003 and 360µm
could be measu ed.
SEM and EDAX analyses we e ca ied ou wi h a Quan a 200 F mic oscope (FEI
Company). The equipmen model 1PW1825 om PHILIPS was used o XRD analysis.
The pa ame e s we e; CuKα1 line (λ= 1.542 A) as adia ion, an X- ay ube po en ial o 40
kV and a cu en o 40 mA.
A ol age and cu en gene a ion sys em based on Field Poin comme cial modules and
LabWindows/CVI so wa e (Na ional Ins umen s, NI) we e used o s udy he in luence o
cu en low on he oxida ion.
Samples wi h po osi ies as low as 30% we e sin e ed o s udy he in luence o a lowe
po osi y on he oxida ion o po ous c o e .
3 RESULTS
3.1 E ec o empe a u e and wa e apou con en
Iso he mal mass gain de e mina ion es s we e pe o med in humidi ied hyd ogen a
empe a u es o 600ºC, 700ºC and 800ªC, and he ollowing wa e apou con en s; 3, 7,
25, 50 and 70%. 3 poin s (2, 24 and 72 hou s) desc ibed he oxida ion kine ics cu es. In
his s udy, only one sample was conside ed enough o de e mine he oxida ion kine ics, as
will be explained la e .
I was ound ha in ai a 800ºC wi h less han 3% o wa e apou ( he o en is open and
ambien humidi y is no con olled bu a ies be ween 2 and 3%), he oxida ion was
ca as ophic, whe eas in a hyd ogen a mosphe e wi h 3% wa e apou con en , he oxide
6
FC-10-1019 An epa a
scale o med is p o ec i e (Fig 1). Highe empe a u es cause highe mass gains. A es o
144 hou s in hyd ogen wi h 3% o wa e apou (H2-3%H2O) was pe o med on one o he
samples in o de o con i m he beha iou o e ime. A SEM analysis (Fig 2 a)) showed
ha om 600 o 800ºC he samples we e oxidized uni o mly in H2-3%H2O. The oxide
laye s we e no hick enough o an EDAX analysis, al hough highe C concen a ions
we e measu ed.
F om he s udy in H2-7%H2O he same conclusions could be d awn, excep ha hicke
oxide laye s a e o med. Thus, he EDAX analysis sugges ed mo e ca ego ically ha he
s oichiome y o he (Fe, C ) spinel was e y close o ha o FeC 2O4.
I was no so clea ha in H2-25%H2O he highe he empe a u e, he highe he mass
gain. A 600, 700 and 800ºC he mass gain a e 2 hou s and 72 hou s was almos he
same, bu always highe han in H2-7%H2O. The concen a ion o Fe in he oxide laye
was highe . Bo h his esul om EDAX and he e y simila mass gains ha we e
measu ed we e a i ied by XRD da a (Fig3 a).
Mass gains wi h a 50% wa e apou con en we e mo e han double he p e ious case
(Fig 4). This can be explained he modynamically because o he o ma ion o di e en i on
oxides depending on he wa e apou con en [2]; a a ce ain empe a u e, as he
pe cen age o wa e apou is g ea e , highe pO2 a e he modynamically in equilib ium. In
H2-25%H2O a 800ºC only C and (C ,Fe) oxides a e he modynamically s able while no
Fe oxide is s able and FeO is s able in H2-50%H2O.
In H2-50%H2O a 700 and 800ºC, he main mass gains occu ed du ing he i s wo
hou s, while 24 hou s we e needed a 600ºC. Only he sample a 800ºC con inued o gain
mass a e he i s wo hou s, up o a mass gain o 35%. This alue is conside ed as he
highes oxida ion s a e o C o e wi h he said pO2 [4]. Pho os om SEM analysis
con i med ha samples a 800ºC we e almos ully oxidized. SEM images om samples in
7
FC-10-1019 An epa a
H2-50%H2O looked like hose in H2-70%H2O (Fig 2 b)). A 600 and 700ºC, non-oxidized
FeC pa icles could be ound easily and he C con en in hese pa icles was lowe han
in C o e . The peak o C o e om XRD da a (Fig3 b)) a ied om being signi ican a e 2
hou s a 600ºC, o ze o a e 72 hou s a 800ºC. In his case, Fe3O4 and (Fe,C )3O4
oxides we e he main phases de ec ed.
Mass gains in H2-70%H2O o all he empe a u es we e a li le highe han in H2-
50%H2O, bu hey p oduced he same cu es. In addi ion, SEM and EDAX analyses we e
e y simila o hose o he samples in H2-50%H2O. Excep ha a 600ºC a e wo hou s,
he XRD peak o C o e was e y weak.
Mass gains a 600ºC we e he lowes among 600ºC, 700ºC and 800ºC o 3 and 7% o
humidi y, and a 700ºC o 25% and highe wa e apou con en s (Fig 4). This can be
explained by di e en i on oxides o med depending on he wa e apou con en and by a
di e en dependence on empe a u e o he di usion coe icien s o Fe and C . Wi h C o e
in ai , high mass gains we e also ob ained a empe a u es o a ound 600ºC. Non-
p o ec i e i on oxides we e de ec ed ou side he c omia a 650ºC [12].
3.2 E ec o cycles.
Two cyclic oxida ion es s we e pe o med in H2-3%H2O; he i s , h ee cycles o 8 hou s
o compa e wi h he iso he mal mass gain a e 24 hou s; he second, ano he h ee cycles
o 24 hou s o compa e wi h he da a a e 72 hou s. The hea ing/cooling slope is
10ºC/min, he ecommended maximum o he op-end alumina ube.
In he igu e showing all he da a wi h 3% apou (Fig. 1), he mass gains we e almos
coinciden . Al hough in he eal case s eepe hea ing/cooling and a la ge numbe o cycles
will ha e o be conside ed, hey seem no o ha e an addi ional de imen al e ec on
oxida ion esis ance.
8
FC-10-1019 An epa a
3.3 E ec o dilu ion in A .
Two kind o dilu ed a mosphe es we e in es iga ed. The same A -10%H2 gas mix was
used in bo h cases, bu he empe a u e o he wa e o he humidi ica ion sys em ba h was
di e en . Based on he assump ion ha he A -10%H2 gas mix was sa u a ed a he ou le
o he humidi ica ion sys em ( he inal dew poin o he gas mix was no measu ed), he
ollowing wo gas composi ions we e ob ained:
- Wi h he empe a u e o he humidi ica ion sys em ba h a 25ºC, i he gas eached
sa u a ion (low lows we e used, so he assump ion o sa u a ion was mo e
easible), absolu e humidi y was 3%. Thus, he inal composi ion was A – 9%H2 –
3%H2O (H2/H2O a io 9/3 = 3)
- Wi h he ba h a 44ºC, he inal composi ion was A – 9%H2 – 9%H2O (H2/H2O
a io 9/9 = 1)
Figu e 5 shows esul s om his oxida ion es and he co ela ion be ween a pu e H2 –
H2O a mosphe e and he same H2/H2O a io dilu ed in A . Lowe kine ics we e ob ained
because he oxidizing low used was no enough. In he end, he same oxida ion le el was
eached, as can be heo e ically expec ed [7]. A e 72 hou s, all samples oxidized in an
a mosphe e wi h he same H2/H2O a io ollowed he same XRD pa e n.
3.4 Use o su ace a ea om Hg po osime y o ep esen mass gain da a.
Di e en echniques we e used o ob ain he su ace a ea; de e mina ion o pe meabili y
[13] and calcula ion o su ace a ea h ough he Kozeny-Ca man equa ion [14], BET, and
Hg po osime y. The bes esul s we e ob ained wi h Hg po osime y.
The samples o he s udy o de e mine he oxida ion kine ics we e cu om a long PM
ube consolida ed a high empe a u e in H2. The pe meabili y o a po ion o his long ube
was measu ed using Hg po osime y, as i is a des uc i e analysis and a e he Hg
po osime y, he p ope ies o he sample changed. The po osime e epo s he su ace
9
FC-10-1019 An epa a
a ea di ec ly. This alue was used o epo he mass gain as a ea-speci ic mass gain and
compa e i wi h he mass gain epo ed as a pe cen age o mass gained om ini ial mass.
As shown in igu e 4, hey looked iden ical and bo h g aphs could be used o epo he
oxida ion kine ics. The mass gain epo ed as a pe cen age o mass gained om ini ial
mass is mo e eliable as i only conside s he e o o wo mass measu emen s. The a ea-
speci ic mass gain has he addi ional e o o he su ace a ea measu emen .
Wi h a single sample, i was no possible o pe o m he Hg po osime y es i s and hen
y o use he sample in he oxida ion es . This is why speci ic mass gains e e ed o he
su ace a ea o a di e en , non-oxidized sample (al hough bo h samples we e cu om he
same ube). Hg po osime y es s we e ca ied ou on oxidized samples, bu u he
in es iga ion is equi ed o each de ini i e conclusions.
The a e age su ace a ea o he samples used in his s udy was 500 imes he
ecommended a ea o dense samples in he mog a ime ic expe imen s [11], which is
why epea abili y was su p isingly so good in his s udy.
3.5 E ec o po osi y.
When conside ing he oxida ion o he me al suppo o an SOFC, i s po osi y is a e y
impo an a iable. In he beginning o he manu ac u ing p ocess o he SOFC cells, he
suppo has a high po osi y, bu i sh inks du ing sin e iza ion. This is why bo h he po osi y
and su ace a ea o he me al-suppo dec ease, as shown in igu e 6. Compa ed o he
speci ic su ace a ea o sphe ical pa icles, i egula ones o he same size ha e a highe
speci ic su ace a ea. The oxida ion o po ous samples depends on speci ic su ace a ea
[2]. Thus, i egula pa icles ha e a weake oxida ion esis ance. The sin e ing p ocess
ans o ms he ini ial shape o he pa icles, i makes he i egula pa icles mo e ound,
and he po osi y and he su ace a ea dec ease a he same ime. These wo aspec s
enhance he oxida ion esis ance o he me al suppo a e sin e ing.
16
FC-10-1019 An epa a
FIGURE CAPTIONS
Fig1. Mass gains a e 2, 24 and 72 hou s a 600ºC, 700ºC and 800ºC. In addi ion, 3
cycles o 8 and 24 hou s in H2-3%H2O a 800ºC.
Fig2. SEM images o samples oxidized a 800ºC o 72h wi h a) 3% o apou , and b) 70%
o apou
Fig3. XRD pa e ns o samples oxidized a 600, 700 and 800ºC o 2, 24 and 72 hou s
wi h a apou con en o a) 25% and b) 50%.
Fig4. Mass gains o he po ous samples a e 72 hou s in H2 wi h apou con en s o 3, 7,
25, 50 y 70% a) e e ed o ini ial mass, and b) e e ed o su ace a ea measu ed
by Hg po osime y be o e oxida ion
Fig5. Mass gains a e 2, 24 and 72 hou s a 600ºC, 700ºC and 800ºC in H2/H2O
a mosphe es ( a ios 1 and 3) dilu ed in A .
Fig6. SEM images o he su ace o h ee non-oxidized samples a) consolida ed a high
empe a u e in hyd ogen (70% po osi y) b) 20% o sh inkage c) 25% o sh inkage
(~30% po osi y)
Fig7. SEM images o samples oxidized a 800ºC in H2/H2O=1 o 70h a) consolida ed a
high empe a u e in hyd ogen (70% po osi y) b) 25% o sh inkage (~30% po osi y)
Fig8. SEM images o samples oxidized a 800ºC in H2/H2O=1 o 70h wi h a cu en low
o a) 0mA/cm2 b) 800mA/cm2
17
FC-10-1019 An epa a
Fig1. Mass gains a e 2, 24 and 72 hou s a 600ºC, 700ºC and 800ºC. In addi ion, 3
cycles o 8 and 24 hou s in H2-3%H2O a 800ºC.
a)
18
FC-10-1019 An epa a
b)
19
FC-10-1019 An epa a
Fig2. SEM images o samples oxidized a 800ºC o 72h wi h a) 3% o apou , and b) 70%
o apou
a)
20
FC-10-1019 An epa a
b)
Fig3. XRD pa e ns o samples oxidized a 600, 700 and 800ºC o 2, 24 and 72 hou s
wi h a apou con en o a) 25% and b) 50%.
a)
21
FC-10-1019 An epa a
b)
22
FC-10-1019 An epa a
Fig4. Mass gains o he po ous samples a e 72 hou s in H2 wi h apou con en s o 3, 7,
25, 50 y 70% a) e e ed o ini ial mass, and b) e e ed o su ace a ea measu ed
by Hg po osime y be o e oxida ion
Fig5. Mass gains a e 2, 24 and 72 hou s a 600ºC, 700ºC and 800ºC in H2/H2O
a mosphe es ( a ios 1 and 3) dilu ed in A .
a)
23
FC-10-1019 An epa a
b)
24
FC-10-1019 An epa a
c)
25
FC-10-1019 An epa a
Fig6. SEM images o he su ace o h ee non-oxidized samples a) consolida ed a high
empe a u e in hyd ogen (70% po osi y) b) 20% o sh inkage c) 25% o sh inkage
(~30% po osi y)
a)
