E ec o a omic subs i u ion on he sodium manganese e i e
he mochemical cycle o hyd ogen p oduc ion
F ancesco To e
a
, Te esa Aguila Sanchez
a
, S e ania Doppiu
a
,
*
,
Mikel O egui Bengoechea
b
, Ped o Luis A ias E gue a
b
, Elena Palomo del Ba io
a
,
c
a
Cen e o Coope a i e Resea ch on Al e na i e Ene gies (CIC ene giGUNE), Basque Resea ch and Technology Alliance (BRTA), Ala a Technology Pa k,
Albe Eins ein 48, 01510 Vi o ia-Gas eiz, Spain
b
Uni e sidad del Pais Vasco - Euskal He iko Unibe si a ea, Escuela de Ingenie ía, Plaza To es Que edo, 1, 48013 Bilbao, Spain
c
Ike basque, Basque Founda ion o Science, Bilbao 348013, Spain
a icle in o
A icle his o y:
Recei ed 8 Ap il 2022
Recei ed in e ised o m
6 July 2022
Accep ed 7 July 2022
A ailable online 25 Augus 2022
Keywo ds:
The mochemical wa e spli ing
Sodium manganese e i e cycle
Ca bona ion
Deca bona ion
Fuel p oduc ion
Ca ion subs i u ion
abs ac
This wo k p esen s he e ec o a omic subs i u ion on he MnFe
2
O
4
eNa
2
CO
3
he mochemical cycle
o H
2
p oduc ion. The non-oxida i e deca bona ion/ca bona ion eac ion o he MnFe
2
O
4
eNa
2
CO
3
mix u e is in es iga ed as he s a ing e e ence. Repea ed cycling esul s in a 30% loss o e e sibili y
due o an o e all educ ion o he eac i e in e aces. The subs i u ion o Na
2
CO
3
o Li
2
CO
3
dec eases
he deca bona ion onse empe a u e by abou 100
C, bu almos no e e sibili y is obse ed du ing
he cycles due o he i e e sible Li
þ
in e cala ion. The e ec o pa ial Mn subs i u ion o Ca, Ni, and
Zn is p esen ed. The 5% Zn mix u e shows he bes deca bona ion/ca bona ion e e sibili y and is
es ed o H
2
p oduc ion oge he wi h MnFe
2
O
4
eNa
2
CO
3
. The e e ence mix u e p oduces mo e H
2
du ing he fi s cycle (z1.1 s. 0.7 mmol/g), bu i s p oduc ion d as ically d ops by wo o de s o
magni ude upon cycling and becomes negligible a e 5 cycles. By con as , he Zn-doped mix u e
exhibi s a s able H
2
p oduc ion o 0.22 mmol/g wi h no dec easing end obse ed om cycle 2 o
cycle 5. As esul , in he fi h cycle, he Zn-doped mix u e p oduces 23 imes mo e H
2
han MnFe
2
O
4
eNa
2
CO
3
. The mog a ime y and X- ay di ac ion confi m ha doping wi h Zn significan ly imp o es
he egene a ion o he eac an s.
©2022 The Au ho (s). Published by Else ie L d. This is an open access a icle unde he CC BY-NC-ND
license (h p://c ea i ecommons.o g/licenses/by-nc-nd/4.0/).
1. In oduc ion
The mochemical wa e spli ing (WS) ep esen s a e y p om-
ising and elegan solu ion o sus ainable la ge-scale hyd ogen
p oduc ion [1e3]. This echnology uses wa e and hea as inpu , o
p oduce H
2
and O
2
. Al hough di ec wa e he molysis can be
achie ed in p inciple, he ex eme empe a u es (T >2200
C) and
he necessi y o sepa a e H
2
om O
2
make his solu ion imp ac ical
[4]. Imp o emen s can be achie ed by di iding he o e all p ocess
in o consecu i e eac ions (minimum wo) ha a e cyclically
epea ed h ough a so-called he mochemical cycle [5]. This allows
o lowe he ope a ing empe a u es and p oduce H
2
and O
2
in
sepa a e s eps, educing he isk o acciden s.
Among mo e han 300 p oposed so a , wo-s ep he mo-
chemical cycles based on me al oxides (e.g. SnO
2
, ZnO, CeO
2
,
Mn
3
O
4
,Fe
x
O
y
) ely on ela i ely low-cos ma e ials and do no
in ol e dange ous o co osi e eac ion in e media es [6e12].
Howe e , he egene a ion ( educ ion) s ep ypically equi es
empe a u es >1500
C, which nega i ely a ec s he long- e m
ma e ials pe o mances and poses significan challenges when
choosing he ma e ials o he he mochemical eac o s [13]. The
addi ion o seconda y compounds has been p oposed o lowe
he eac ion empe a u es below 1000
C. In his ega d, a
ecen assessmen conside ing p ocess economics, en i on-
men al impac , cyclabili y, and simplici y o ope a ion high-
ligh ed he MnFe
2
O
4
eNa
2
CO
3
cycle as one o he bes o
p ac ical applica ions [14]. This cycle was ini ially p oposed by Y.
Tamamu a e al. [15,16] and hen u he in es iga ed by C.
Al ani e al. [17e25]. Beyond lab-scale expe imen s, he easi-
bili y o his he mochemical cycle has also been p o en in a
small sola concen a ion acili y [25].
In fi s app oxima ion, his he mochemical cycle consis s o
wo s eps [15] and can wo k a empe a u es a ound 800e750
C:
*Co esponding au ho .
E-mail add ess: [email p o ec ed] (S. Doppiu).
Con en s lis s a ailable a ScienceDi ec
Ma e ials Today Ene gy
jou nal homepage: www.jou nals.else ie .com/ma e ials- oday-ene gy/
h ps://doi.o g/10.1016/j.m ene .2022.101094
2468-6069/©2022 The Au ho (s). Published by Else ie L d. This is an open access a icle unde he CC BY-NC-ND license (h p://c ea i ecommons.o g/licenses/by-nc-nd/4.0/).
Ma e ials Today Ene gy 29 (2022) 101094
2MnFe2O4ðsÞþ3Na2CO3ðsÞþH2OðgÞ/6NaFe
2
=
3
Mn
1
=
3
O2ðsÞ
þH2ðgÞþ3CO2ðgÞ
(1)
6NaFe
2
=
3
Mn
1
=
3
O2ðsÞþ3CO2ðgÞ/2MnFe2O4ðsÞþ3Na2CO3ðsÞ
þ0:5O2ðgÞ
(2)
Howe e , a mo e complex eac ion pa h was expe imen ally
obse ed, as i was clea ly desc ibed by Va sano e al. [23]. In
pa icula , he hyd ogen p oduc ion eac ion (1) was shown o
p oceed h ough wo sepa a e s eps. Fi s , he non-oxida i e pa ial
deca bona ion akes place, whe e 2/3 o Na
2
CO
3
eac s wi h
MnFe
2
O
4
. As a esul o his eac ion, CO
2
is eleased, and sodium
in e cala es in he cubic spinel phase o o m a fine mix u e o
NaFeO
2
and MnO:
2MnFe2O4ðsÞþ3Na2CO3ðsÞ/2ðMnO*2NaFeO2ÞðsÞþNa2CO3ðsÞ
þ2CO2ðgÞ
(3)
The o ma ion o he (MnO*NaFeO
2
) compound pa es he way
o he subsequen eac ion wi h H
2
O s eam o p oduce H
2
(4). Mn
2þ
is oxidized o Mn
3þ
o o m NaFe
2/3
Mn
2/3
O
2
, which has he same
c ys al s uc u e as NaFeO
2
. A his poin , mo e Na
þ
can be in e -
cala ed, which d i es he decomposi ion o he emaining Na
2
CO
3
.
2ðMgO*2NaFeO2ÞðsÞþNa2CO3ðsÞþH2OðgÞ/6NaMn
1
3
Fe
2
3
O2ðsÞþ2CO2ðgÞþH2ðgÞ(4)
Once he WS eac ion is comple ed, NaFe
2/3
Mn
2/3
O
2
is exposed
o a CO
2
- ich gas, which p omo es he sodium oxide dein e cala ion
o o m Na
2
CO
3
. As esul , he Na con en in he NaFe
2/3
Mn
2/3
O
2
p og essi ely dec eases and he laye ed s uc u e collapses. The
educ ion o Mn
3þ
o Mn
2þ
and he co esponding elease o mo-
lecula oxygen e en ually leads o he egene a ion o MnFe
2
O
4
and
Na
2
CO
3
:
6NaFe
2
=
3
Mn
1
=
3
O2ðsÞþ3CO2ðgÞ/2MnFe2O4ðsÞþ3Na2CO3ðsÞ
þ1
2O2ðgÞ
(5)
Un o una ely, pa icle sin e ing and coalescence o a Na
2
CO
3
laye end o hinde he comple e egene a ion o he s a ing e-
ac an s, leading o a d as ic dec ease in he H
2
p oduc ion upon
p olonged cycling [24].
In gene al, he imp o emen o he eac ion kine ics and
e e sibili y, as well as he dec ease o he ope a ing empe a u es,
is highly impo an o inc ease he e ficiency o he mochemical
cycles, as well as o b oaden hei po en ial fields o applica ion
[26]. Speed up he eac ion kine ics means inc easing he H
2
p o-
duc ion pe uni o ime, while long- e m s abili y o he ma e ials
is undamen al o he o e all p ocess o be easible and compe i-
i e. In his sense, lowe ing he ope a ing empe a u e educes he
sin e ing o he used oxides and educes he equi emen s o he
ma e ials needed o cons uc he mochemical eac o s.
A omic subs i u ion, educ ion o he c ys alli e size, and addi-
ion o ine compounds a e common s a egies o imp o e ma e-
ials pe o mances [7,27e35]. Fo ins ance, a omic subs i u ion
significan ly imp o ed he pe o mances o CeO
2
- and pe o ski es-
based he mochemical cycles o uel p oduc ion [36e41]. A
s aigh o wa d ela ion be ween cycle pe o mances and he
mic os uc u al changes induced by a omic subs i u ion was e-
po ed in some cases. The egene a ion empe a u e o some doped
pe o ski es dec eased when in oducing doping elemen s wi h
lowe a omic size. Such beha io was a ibu ed o an inc ease in
he a omic size misma ch ha , in u n, led o a highe local diso de
[38]. Conce ning e i e-ca bona e cycles, in es iga ions pe o med
on single me al oxides (Mn
3
O
4
,Fe
3
O
4
, and Co
3
O
4
) and ca bona es
(Na
2
CO
3
,Li
2
CO
3
, and K
2
CO
3
) showed ha he H
2
e olu ion a e was
ound o a y depending on he specific ca ion combina ions [42].
To he bes o ou knowledge, simila app oaches ha e no ye
been epo ed o he MnFe
2
O
4
eNa
2
CO
3
he mochemical cycle.
Thus, he p esen wo k ep esen s he fi s a emp in his di ec-
ion. Pa icula a en ion is gi en o he non-oxida i e deca bon-
a ion eac ion ( eac ion 3) as i ep esen s he fi s s ep o he
o e all he mochemical cycle. The deca bona ion-ca bona ion e-
ac ion o MnFe
2
O
4
eNa
2
CO
3
is ollowed du ing 10 consecu i e cy-
cles, and he ela ed s uc u al changes a e in es iga ed. The e ec
o a omic subs i u ion is hen explo ed by using MnFe
2
O
4
eNa
2
CO
3
as he basis o compa ison. S a ing om he MnFe
2
O
4
eNa
2
CO
3
sys em, he e ec o a omic subs i u ion on i s eac i i y is aced
h ough wo dis inc app oaches. E alua ing he subs i u ion o Na
o Li and K ca ions and pa ially subs i u ing Mn
2þ
o Ca
2þ
,Ni
2þ
,
and Zn
2þ
. In he la e case, e i es wi h gene al o mula Mn
1-
x
A
x
Fe
2
O
4
, wi h A being Ni, Zn, Ca, and x ¼0.05, 0.10, and 0.50 a e
syn hesized and es ed h ough deca bona ion/ca bona ion cycles
oge he wi h Na
2
CO
3
. The mos p omising doped mix u e is hen
es ed o he H
2
p oduc ion eac ion, and i s pe o mances a e
compa ed wi h hose o he e e ence MnFe
2
O
4
eNa
2
CO
3
sys em.
2. Ma e ial and me hods
2.1. Syn hesis o Mn e i e oxides
Manganese e i e-based oxides, wi h spinel s uc u e, we e
syn hesized ollowing he sel -combus ion me hod; glycine
(C₂H₅NO₂, 98.5%, Sigma Ald ich) was used as complexan and uel
agen [43]. An aqueous solu ion o Fe(NO
3
)
3
$9H
2
O (99%, Al a Aesa )
and Mn(NO
3
)
2
$4H
2
O (99%, Al a Aesa ) was p epa ed; a mola a io
o 2:1 be ween Fe
3þ
and Mn
2þ
and a o al ca ion concen a ion o
1 M we e ensu ed. A 1 M solu ion o glycine was hen added and
he final solu ion was s i ed o 2 h. The solu ion was ans e ed
in o a Pe i- ype dish and kep a 100
C un il a ed-b ownish gel
was o med. Finally, he sel -combus ion eac ion was ca ied ou a
350
C, whe e a black-b ownish, oamy p oduc was ob ained.
The same me hodology was used o syn hesize modified manga-
nese e i es wi h gene al o mula Mn
1-x
A
x
Fe
2
O
4
, wi h A being Ni, Zn,
Ca, and x ¼0.05, 0.1, and 0.5. To his end, pa o he Mn
2þ
p ecu so
was subs i u ed by he ni a e o he co esponding di alen dopan
ca iondi.e. Zn(NO
3
)
2
$6H
2
O (98%, Al a Aesa ), Ni(NO
3
)
2
$6H
2
O
(Essen Q®, Scha lau), and Ca(NO
3
)
2
$4H
2
O(99%,SigmaAld ich).On
he o he side, he Fe
3þ
concen a ion was no changed. In his way,
he a io be ween he di alen and i alen ca ions was main ained
cons an a 1:2. The a omic dis ibu ion o di alen and i alen
ca ionsin he oc ahed al and e ahed al si es o he spinel la ice can
change depending on he specific chemical composi ion, he syn-
hesis me hod, and he mic os uc u al ea u es [44,45]. The de e -
mina ion o he exac a omic dis ibu ion is no s aigh o wa d and
willno beadd essedin hiswo k.In ac , heMnsubs i u ionwe e e
o should be in ended only in e ms o chemical composi iondi.e. he
pa ial eplacemen o Mn
2þ
wi h Ca
2þ
,Ni
2þ
,andZn
2þ
.
Fo he sake o simplici y, he di e en oxides will be indica ed
by e e ing o he pe cen age o Mn a omic subs i u ion; o
F. To e, T.A. Sanchez, S. Doppiu e al. Ma e ials Today Ene gy 29 (2022) 101094
2
ins ance, Mn
0.5
Ca
0.5
Fe
2
O
4
will be e e ed o as 50 a .% Ca,
Mn
0.95
Ni
0.05
Fe
2
O
4
as 5 a .% Ni, e c. (see Table 1).
2.2. The mal analysis
2.2.1. Non-oxida i e deca bona ion eac ion o he
MnFe
2
O
4
eNa
2
CO
3
mix u e
Tempe a u e p og ammed deso p ion (TPD) expe imen s we e
pe o med o in es iga e he deca bona ion/ca bona ion beha io
o he di e en mix u es o spinel e i es and ca bona es. Fo each
expe imen , he selec ed spinel e i e and ca bona e we e mixed a
a 1:1 M a io using a mo a . The choice o his s oichiome y can be
unde s ood by looking a he non-oxida i e deca bona ion s ep
desc ibed by Eq. (3). In non-oxidizing condi ions (absence o H
2
O),
only 2/3 o he Na
2
CO
3
eac s wi h MnFe
2
O
4
, while he emaining 1/
3o Na
2
CO
3
does no pa icipa e in he eac ion. This co esponds o
a MnFe
2
O
4
:Na
2
CO
3
mola a io o 1:1. Abou 15e20 mg o he
mix u e we e ans e ed o an alumina c ucible and loca ed in a
NETZSCH STA 449 F3 he mobalance. P elimina y iso he mal ex-
pe imen s we e pe o med on he MnFe
2
O
4
eNa
2
CO
3
mix u e a
600, 700, 750, 800, and 850
C. The samples we e hea ed unde N
2
a 10
C/min o each he desi ed empe a u e and kep in
iso he mal condi ions un il he eac ion was comple ed ( he
du a ion o he iso he mal s ep a ies depending on he empe a-
u e). Du ing hese expe imen s, he CO
2
elease was analyzed by
using a mass spec ome e (NETZSCH QMS 403C).
The e e sibili y o he non-oxida i e deca bona ion-ca bon-
a ion was hen in es iga ed h ough deca bona ion-ca bona ion
cycles. The empe a u e was inc eased om oom empe a u e o
800
C a a cons an hea ing a e o 10
C$min
1
unde N
2
and kep
cons an o 30 min in N
2
o induce he comple e elease o he CO
2
.
The ca bona ion p ocess was ca ied ou unde CO
2
a mosphe e
wi h a flow o 50 mL$min
1
dec easing he empe a u e o 400
C
a a cons an cooling a e o 10
C$min
1
. Cycling expe imen s we e
also pe o med on he MnFe
2
O
4
eNa
2
CO
3
1:1 mix u e. To his aim,
he empe a u e was again inc eased o 800
C, and he empe a-
u e p og am was epea ed depending on he desi ed numbe o
cycles. A maximum o 10 cycles was pe o med. The same me h-
odology was subsequen ly used o es he e ec o Mn a omic
subs i u ion. The syn hesized oxides wi h 5 and 10 a .% o Ca, Ni,
and Zn (see Sec ion 2.1) we e mixed wi h Na
2
CO
3
a a 1:1 M a io.
The oxides wi h 50 a .% o Mn subs i u ion we e excluded due o
he p esence o seconda y phases (see s uc u al cha ac e iza ion in
he Suppo ing in o ma ion).
The subs i u ion o Na
2
CO
3
o Li
2
CO
3
and K
2
CO
3
on eac ion 3
was also aken in o conside a ion. The deca bona ion/ca bona ion
o equimola MnFe
2
O
4
eLi
2
CO
3
and MnFe
2
O
4
eK
2
CO
3
mix u es was
in es iga ed by cycling expe imen s. Hea ing was pe o med a
10
C/min unde N
2
(50 mL$min
1
); di e en empe a u es we e
es ed wi h a maximum o 925
C. A e a sho iso he m o 10 min,
he sample was cooled down o 400
Ca 10
C/min unde 50 mL/
min o CO
2
. The p og am was hen epea ed wo mo e imes o ge
h ee deca bona ion/ca bona ion cycles.
2.2.2. Hyd ogen p oduc ion
H
2
p oduc ion expe imen s we e pe o med using an STA 449
F3 Jupi e (NETZSCH) he mobalance coupled wi h a wa e apo
gene a o p o ided byaDROP GmbH. The ou le gas o he STA o en
was connec ed o an H
2
Cla k- ype mic osenso (de ec ion
limi z10
2
ol%) in e aced wi h an amplifie (Unisense,
Denma k). The senso was calib a ed be o e each measu emen by
using a wo-poin calib a ion as ecommended by he p o ide . To
his aim, pu e A and an A wi h 2 ol% H
2
s anda ds we e used. Fo
each expe imen , he desi ed spinel was mixed wi h Na
2
CO
3
a a
2:3 a io (see eac ion 1) using a mo a and hen p essed a 1 on
o 30 s o ob ain a 400 mg pelle (d ¼10 mm and h z2 mm).
Despi e he use o powde s would ha e ensu ed highe gas-solid
in e aces, he use o pelle s was necessa y o wo k wi h a sample
size la ge enough o allow H
2
quan ifica ion. The H
2
p oduc ion
expe imen s we e pe o med as ollows. The empe a u e was
inc eased o 800
C(10
C$min
1
) and kep cons an o 30 min o
elease 2/3 o he o al CO
2
( eac ion 3); only A flowed du ing his
fi s pa (30 mL/min). Wa e apo was hen in oduced o 90 min
a a a e o 1 g H
2
O/h while keeping he A flow a 30 mL/min. The
wa e supply was hen s opped, and he A flow was main ained o
10 min o emo e he emaining H
2
O. CO
2
was hen in oduced
(60 mL/min) and he sample cooled a 10
C/min. When he em-
pe a u e eached 400
C, he gas was hen shi ed o A and he
empe a u e inc eased again o 800
C o s a a new cycle. A o al
o 10 cycles we e pe o med. Addi ional expe imen s we e pe -
o med in iso he mal condi ions a 750
C and using an H
2
mic o-
senso wi h a lowe de ec ion limi (z10
3
ol%; Unisense,
Denma k). The samples we e hea ed a 10
C/min o 750
C unde a
100 mL N
2
flow. Wa e apo was hen in oduced in o he gas flow
a a a e o 0.5 g H
2
O/h o 5 h. When he wa e supply was s opped,
he N
2
flow was main ained o 5 min o emo e he emaining
H
2
O. CO
2
was hen in oduced (100 mL/min) and he sample was
le a 750
C o e nigh . The gas flow was hen shi ed again o s a
ano he WS s ep. A o al o 5 cycles we e pe o med. As he H
2
p oduc ion was obse ed o be as e om he second cycle, he 4
emaining WS s eps we e sho ened om 5 o 3 h.
2.3. S uc u al cha ac e iza ion
The as-syn he ized e i es and he s udied mix u es we e
cha ac e ized by X- ay di ac ion (XRD) analysis using a B uke D8
Disco e equipped wi h a LYNXEYE XE de ec o and a mono-
ch oma ic Cu K
a
1 adia ion sou ce o
l
¼1.54056 Å. Fo some
mix u es, a sample holde equipped wi h a Kap on film was used o
a oid oxida ion and/o hyd a ion du ing he XRD pa e n acquisi-
ion. Fo ins ance, MnO is e y p one o oxida ion o Mn
3þ
species
and d y K
2
CO
3
apidly o ms sesquihyd a e K
2
CO
3
. The use o
Kap on esul s in a s ong backg ound a low sca e ing angles
(10 >2
q
<30). In si u XRD measu emen s we e pe o med a
di e en empe a u es using a B uke D8 Ad ance di ac ome e
equipped wi h a LYNXEYE de ec o , a Van ec-1 PSD de ec o , a Co
ube (K
a
1o
l
¼1.78896; K
b
1¼1.6210334), and an An on Pa
HTK2000 high- empe a u e u nace. Pa e ns we e eco ded unde
N
2
a mosphe e a oom empe a u e and 400, 450, 500, 550, 600,
700, and 800
C in a s epwise manne . The deca bona ion eac ion
was analyzed as ollows. A e a s abiliza ion ime o 30 min a
800
C, CO
2
was hen in oduced, and he empe a u e was hen
main ained a 800
C o 1 h o gua an ee a comple e sa u a ion o
he chambe . Pa e ns we e hen collec ed a 800, 600
C, and oom
empe a u e in a s epwise manne wi h a cooling a e o 10
C/min
a e each s ep. Phase iden ifica ion was pe o med using he EVA
Table 1
Composi ions o he di e en syn hesized oxides oge he wi h he abb e ia ion
used in he ex .
Composi ion Abb e ia ion
Mn
0.95
Ni
0.05
Fe
2
O
4
5% Ni
Mn
0.9
Ni
0.1
Fe
2
O
4
10% Ni
Mn
0.5
Ni
0.5
Fe
2
O
4
50% Ni
Mn
0.95
Zn
0.05
Fe
2
O
4
5% Zn
Mn
0.9
Zn
0.1
Fe
2
O
4
10% Zn
Mn
0.5
Zn
0.5
Fe
2
O
4
50% Zn
Mn
0.95
Ca
0.05
Fe
2
O
4
5% Ca
Mn
0.9
Ca
0.1
Fe
2
O
4
10% Ca
Mn
0.5
Ca
0.5
Fe
2
O
4
50% Ca
F. To e, T.A. Sanchez, S. Doppiu e al. Ma e ials Today Ene gy 29 (2022) 101094
3
di ac ion comme cial so wa e. The collec ed pa e ns we e hen
analyzed acco ding o he Rie eld me hod [46] by using he MAUD
so wa e [47]. The mo phology o he powde s be o e and a e
ca bona ion-deca bona ion cycles was analyzed using scanning
elec on mic oscopy (SEM) using a Quan a 200 FEG (FEI Company,
Hillsbo o, OR, USA) ope a ing in high acuum mode.
3. Resul s and discussion
3.1. Non-oxida i e deca bona ion eac ion o he MnFe
2
O
4
eNa
2
CO
3
mix u e
The mo phology o he as-syn hesized oxides was in es iga ed
using SEM. Rep esen a i e images o he MnFe
2
O
4
a e epo ed in
Fig. 1aeb. The ob ained powde s show a hie a chical ne wo k wi h
well-connec ed mac opo es due o he consis en amoun o gases
ha is apidly eleased du ing he eac ion [48]. The syn hesized
e i es we e cha ac e ized using XRD, and he phase composi ion
and he mic os uc u e we e in es iga ed acco ding o he Rie eld
me hod (Sec ion 1 o Suppo ing in o ma ion). The ob ained pow-
de s p esen a fine mic os uc u e wi h he a e age c ys alli e sizes
in he ange o 50e100 nm, which is due o he as cooling expe-
ienced du ing he combus ion syn hesis. In all cases, he main
c ys allog aphic phase is he cubic spinel cha ac e is ic o he
MnFe
2
O
4
compound, ollowed by mino amoun s o FeO and MnO
(Fig. S1a). Mino amoun s o o he seconda y phases we e de ec ed
in he samples wi h 50% o a omic subs i u ion. The a omic sub-
s i u ion also a ec ed he la ice pa ame e o he cubic spinel
phase (Fig. S1a). Di e en ends we e obse ed depending on he
a omic misma ch be ween he doping elemen and Mn. While Zn
did no induce any significan changes, Ca and Ni led o la ice
expansion and sh inkage, espec i ely.
The non-oxida i e deca bona ion eac ion (3) o he
MnFe
2
O
4
eNa
2
CO
3
sys em was in es iga ed unde iso he mal con-
di ions in he 600e850
C empe a u e ange. The TPD p ofiles
ob ained o he 1:1 mix u e a inc easing empe a u es a e e-
po ed in Fig. 1c. Fo all he expe imen s, a small mass dec ease
(z0.5 w %) is de ec ed a empe a u es a ound 100
C because o
he deso p ion o abso bed wa e . The decomposi ion o he
mix u e s a s a app oxima ely 550
C and p oceeds a di e en
a es depending on he empe a u e iso he m. As expec ed om
eac ion 3, mass spec ome (MS) analysis o he ou le gases
confi ms ha CO
2
was he only p oduc . The in ensi y o he
cha ac e is ic peak o he CO
2
þ
ion (44 m/z) is epo ed in Fig.1dasa
unc ion o ime. Fo he sample hea ed a 600
C, he deca bon-
a ion eac ion p oceeds e y slowly and only a small mass change
(z2 w %) is de ec ed a e mo e han 300 min. No significan
changes can be obse ed in he MS signal apa om a small peak
be ween 25 and 30 min. As he eleased CO
2
is dilu ed in he N
2
ca ie gas, i is possible ha he esul ing concen a ion was oo
low o be de ec ed. The eac ion p oceeds as e a 700 and 750
C.
A significan mass loss is obse ed, oge he wi h an inc ease o he
MS signal ha eaches a maximum a e 33 and 36 min, espec-
i ely. Then, he CO
2
e olu ion con inues a a lowe a e and he
slope o he mass change p ofile p og essi ely dec eases. A e
180 min a 700
C, a mass loss o abou 10 w % is obse ed, meaning
ha he ull deso p ion could no be achie ed wi hin a easonable
ime a his empe a u e. On he o he hand, a pla eau is eached
Fig. 1. (a), (b) SEM images o he as-syn hesized MnFe
2
O
4
a di e en magnifica ions. (c) TPD p ofiles o he MnFe
2
O
4
eNa
2
CO
3
mix u e (1:1) in he 600e850 C empe a u e ange.
(d) MS signal o he CO
2
þ
ion (m/z ¼44) de ec ed du ing he iso he mal expe imen s. (e) In si u XRD pa e ns o he MnFe
2
O
4
eNa
2
CO
3
mix u e (1:1) upon hea ing om RT o 800 C
unde N
2
. SEM, scanning elec on mic oscopy; TPD, empe a u e p og ammed deso p ion; XRD, X- ay di ac ion.
F. To e, T.A. Sanchez, S. Doppiu e al. Ma e ials Today Ene gy 29 (2022) 101094
4
a e abou 60 min a 750
C. The o e all mass loss is a ound 12 w %,
which acco ding o eac ion 3 is close o he heo e ical one (13 w
%). The mix u es hea ed a 800 and 850
C ollow he p e ious
end. The maximum CO
2
concen a ion is e ealed a e 38 and
40 min, while deso p ion ends a e 85 and 60 min, espec i ely. In
bo h cases, he mass s abilizes a abou 87.5 w %, hus sugges ing
ha he eac ion almos goes o comple ion.
XRD analysis o he MnFe
2
O
4
eNa
2
CO
3
mix u e a e he TPD a
850
C is in line wi h he p e ious obse a ion (Fig. S2a). Abou
70 w % and 25 w % we e ob ained o NaFeO
2
and MnO, espec-
i ely, oge he wi h a 5 w % o un eac ed spinel. This ag ees wi h
he heo e ical alues calcula ed based on eac ion 3 (75.5 w % and
24.5 w % o NaFeO
2
and MnO, espec i ely) p oposed by Va sano
e al. [23]. Howe e , he same au ho s did no obse e he o ma-
ion o MnO due o he highly dispe sed and diso de ed na u e o
he manganese compound. To ob ain u he insigh , XRD pa e ns
we e acqui ed in si u upon hea ing he MnFe
2
O
4
eNa
2
CO
3
unde an
N
2
a mosphe e Fig. 1e. A significan inc ease in he la ice pa am-
e e o he spinel phase can be app ecia ed due o he mal expan-
sion (Fig. S2b). In addi ion, he ansi ion o
g
-Na
2
CO
3
o
a
-Na
2
CO
3
akes place a 400
C acco ding o wha was p e iously epo ed
[23]. A 600
C, he NaFeO
2
s a s o ming as highligh ed by he
peaks a 2
q
alues o 19 and 47, while MnO is no de ec ed p obably
due o i s low amoun . A 700
C, he amoun o NaFeO
2
inc eases
while he in ensi y o he B agg peaks o MnFe
2
O
4
significan ly
educes. Mo eo e , a small peak o he MnO can be app eciable a
2
q
¼47. Finally, he pa e n collec ed a 800
C confi ms ha e-
ac ion 3 wen o comple ion, as only NaFeO
2
and MgO a e de ec ed.
The cyclabili y o he MnFe
2
O
4
eNa
2
CO
3
mix u e was in es i-
ga ed h ough 10 consecu i e deca bona ion/ca bona ion cycles.
An equilib a ion cycle was ini ially pe o med o educe any di -
e ence among he a ious ba ches o syn hesized spinel. The
sample was ini ially hea ed up o 800
C, and his empe a u e was
kep cons an o 2 h. Cooling o 400
C was hen ca ied ou unde
CO
2
. This cycle will be e e ed o as cycle 0. Ten addi ional cycles
we e hen pe o med in he 400e800
C empe a u e ange. The
mass p ofile and he mass changes measu ed du ing he deca -
bona ion and ca bona ion s eps a e epo ed in Fig. 2a and b,
espec i ely. Du ing cycle 0, almos ull deso p ion is achie ed a e
abou 30 min a 800
C, wi h an o e all mass loss o a ound 11.5 w
%(Fig. 2b). A 10 w % mass inc ease is obse ed du ing he subse-
quen ca bona ion s ep, hus ma king a e e sibili y loss o abou
1.5 w %. A small bu cons an loss o e e sibili y is obse ed om
cycle 1 o cycle 4, while he mass change s abilizes a a alue a ound
8 w % o he ini ial mass du ing he ollowing cycles. SEM analysis
was ca ied ou a e he hi d deso p ion s ep and a e h ee ull
cycles, i.e., a e 455 and 485 min (Fig. 2ced and Fig. 2ee ,
espec i ely). Pa ial sin e ing occu ed due o he p olonged
exposu e o high empe a u es, and he sample showed g ains
anging om 0.5 o 4
m
m wi h an a e age alue o z1.3
m
m(Fig. 2c).
Despi e his, he deso bed sample s ill p esen s a significan
po osi y. A close look un a els he p esence o a lamina s uc u e
ha can be a ibu ed o he NaFeO
2
phase (Fig. 2d) [24]. The
sample a e he ca bona ion s ep shows a mo e i egula
mo phology (Fig. 2ee ). The backsca e ed elec on de ec o high-
ligh s he o ma ion o Na
2
CO
3
(b igh e a eas) ha esul s in high
con as wi h he unde lying MnFe
2
O
4
/NaFeO
2
ma ix. XRD anal-
ysis was pe o med a e 10 cycles wi hou showing any pa asi e
phases (Fig. S2 ). Howe e , he quan ifica ion o he c ys allo-
g aphic phases sugges s ha a e 10 cycles almos hal o he ini ial
spinel phase do no pa icipa e in he eac ion, and a he ac s as an
ine phase. Based on hese esul s, he loss o e e sibili y can be
explained as ollows (see Fig. 2g). As he ca bona ion eac ion akes
place and he Na ion a e dein e cala ed om NaFeO
2
, he egen-
e a ed Na
2
CO
3
ends o coalesce and co e s he newly o med
MnFe
2
O
4
. The specific in e aces be ween MnFe
2
O
4
and Na
2
CO
3
and
CO
2
a e hen educed, and he subsequen deca bona ion eac ion
is nega i ely a ec ed. In his sense, phenomena like pa ial sin-
e ing, loss o po osi y, and g ain g ow h may u he boos his
beha io , which explains he loss o e e sibili y highligh ed by
he mog a ime y.
3.2. E ec o di e en alkali ca bona es
In his sec ion, we p esen he e ec o he subs i u ion o Na o
K and Li on he kine ic and e e sibili y o he deca bona ion-
ca bona ion eac ion. A o al o h ee deca bona ion-ca bona ion
cycles we e su ficien in his case o highligh he di e ence be-
ween he h ee composi ions. The mass p ofiles o he
MnFe
2
O
4
eK
2
CO
3
and MnFe
2
O
4
eLi
2
CO
3
mix u es (1:1) a e
compa ed wi h he MnFe
2
O
4
eNa
2
CO
3
e e ence sys em (Fig. 3aeb).
The XRD pa e ns acqui ed o he h ee mix u es be o e and a e
he h ee cycles a e epo ed in Fig. 3c. As he beha io o he
MnFe
2
O
4
eNa
2
CO
3
mix u e has been ex ensi ely discussed in he
p e ious sec ion, i will no be add essed in he p esen one and
will be only used as he e m o compa ison o desc ibe he e ec o
Na subs i u ion o Li and K.
The MnFe
2
O
4
eK
2
CO
3
mix u e shows a simila onse empe a u e
o he MnFe
2
O
4
eNa
2
CO
3
one, bu he o e all deca bona ion kine ic is
slowe han he e e ence sys em (Fig. 3a). Du ing he fi s deso p ion
s ep, he mix u e wi h K
2
CO
3
shows a 10.8% mass loss, which is
sligh ly lowe han he heo e ical one (11.9 w %). As CO
2
is in o-
duced, he ca bona ion eac ion immedia ely s a s and 96.5 w % o
he ini ial mass is achie ed in less han 1 min. A dec ease in pe o -
mance is obse ed du ing he ollowing wo cycles, and he deso p-
ion kine ics o MnFe
2
O
4
eK
2
CO
3
u he slows down compa ed o
MnFe
2
O
4
eNa
2
CO
3
. Mo eo e , he sys em loses e e sibili y upon
cycling and deso bs 9.38% and 8.73 w % o i s ini ial mass du ing he
second and hi d cycle, espec i ely. The XRD analysis o he
MnFe
2
O
4
eK
2
CO
3
mix u e a e h ee cycles confi ms he pa ial
egene a ion o he s a ing eac an s, meaning MnFe
2
O
4
and K
2
CO
3
sesquihyd a e (COD 2200592). The la e was de ec ed also in he
s a ing mix u e and is due o he highly hyg oscopic na u e o
anhyd ous K
2
CO
3
, which causes i s apid hyd a ion du ing he XRD
sample p epa a ion. Mo eo e , he cycled MnFe
2
O
4
eK
2
CO
3
mix u e
shows a significan amoun o po assium be a e i e, K
2
Fe
10
O
16
(COD
1529668), which explains he loss o e e sibili y obse ed upon
cycling. The o ma ion o KFeO
2
would be expec ed i he
MnFe
2
O
4
eK
2
CO
3
ollowed eac ion 3. Howe e , no KFeO
2
could be
de ec ed, which sugges s a di e en eac ion mechanism compa ed
o he MnFe
2
O
4
eNa
2
CO
3
sys em. I iswo hsaying ha hepo assium
be a e i e shows a lamella s uc u e ha allows o a b oad ange o
non-s oichiome ic composi ions co esponding o di e en Fe o K
a ios [49]. The compensa ion o non-s oichiome y can be achie ed
by he inclusion o bi alen a oms, which in his case may be Mn
2þ
[50]. Fo hese easons, he a ionaliza ion o hese esul s o find a
eac ion mechanism explaining he loss o pe o mances obse ed in
he MnFe
2
O
4
eK
2
CO
3
mix u e is no s aigh o wa d.
The MnFe
2
O
4
eLi
2
CO
3
sys em demons a es a di e en beha io .
Du ing he fi s hea ing s ep, he sys em shows imp essi e pe -
o mances in e ms o kine ics and shows an onse deca bona ion
empe a u e o abou 100
C lowe han he MnFe
2
O
4
eNa
2
CO
3
mix u e (Fig. 3a). I s a s eleasing CO
2
a a ound 420
C and
almos comple es he deca bona ion when he MnFe
2
O
4
eNa
2
CO
3
and MnFe
2
O
4
eK
2
CO
3
mix u es s a deca bonizing. The obse ed
o e all mass loss is abou 11.37 w %, which is 78% o he 14.4 w %
heo e ical alue. Howe e , he subsequen exposu e o CO
2
does
no induce any significan mass inc ease and he mass o he
sample emains almos un a ied du ing he subsequen wo cycles
(Fig. 3b). Such poo e e sibili y o he MnFe
2
O
4
eLi
2
CO
3
sys em is
F. To e, T.A. Sanchez, S. Doppiu e al. Ma e ials Today Ene gy 29 (2022) 101094
5
confi med by he XRD analysis (Fig. 3c), which highligh s he
p esence o MnFe
2
O
4
and cubic LiFeO
2
(COD 1541312). Thus, i
seems ha he eac ion ollows he mechanism o eac ion 3, wi h
LiFeO
2
being o med ins ead o NaFeO
2
. Howe e , he subsequen
dein e cala ion o Li ions seems o be hinde ed p obably due o he
s ong LieO bond. As esul , eac ion 3 is i e e sible a he p esen
expe imen al condi ions o he MnFe
2
O
4
eLi
2
CO
3
sys em. The XRD
quan i a i e analysis indica es 49 w % and 51 w % o MnFe
2
O
4
and
LiFeO
2
, espec i ely. Conside ing he mola mass o he wo com-
pounds, his means ha he mola a io be ween MnFe
2
O
4
and
LiFeO
2
is a ound 1:2.5. I he s oichiome y o eac ion 3 is alid also
o he MnFe
2
O
4
eLi
2
CO
3
sys em, his means ha only abou 55% o
he ini ial mix u e eac ed. This would imply a mass loss o only
7.9 w %, which is lowe han he expe imen ally obse ed and
sugges s ha some c oss- eac ion may ha e aken place. Mo eo e ,
he spinel phase in he MnFe
2
O
4
eLi
2
CO
3
cycled mix u e has an
a e age la ice pa ame e o 8.38 Å, which is significan ly lowe
han he one obse ed o he MnFe
2
O
4
eNa
2
CO
3
cycled in he same
expe imen al condi ions (8.53 Å). This has a leas wo possible
explana ions. Fi s , Fe
3
O
4
is o med a he han MnFe
2
O
4
; indeed,
he wo compounds sha e he same c ys al s uc u e wi h he only
di e ence being he la ge la ice pa ame e o MnFe
2
O
4
as a
consequence o Fe
2þ
subs i u ions o he bigge Mn
2þ
. Howe e ,
he o ma ion o Fe
3
O
4
would imply he educ ion o pa o Fe
3þ
o
Fe
2þ
a he expenses o Mn
2þ
ha should be oxidized o o m
NaMnO
2
o o he seconda y compounds, which we e no de ec ed.
Ano he possibili y is ha he MnFe
2
O
4
la ice has sh unk due o a
pa ial Li
þ
subs i u ion o in e cala ion, as he a omic subs i u ion
o low elec onega i e ca ions was epo ed o induce a dec ease in
spinel la ice pa ame e s [51e53].
Resuming, he MnFe
2
O
4
eNa
2
CO
3
mix u e showed be e pe -
o mances han bo h MnFe
2
O
4
eLi
2
CO
3
and MnFe
2
O
4
eK
2
CO
3
.
Indeed, he o me showed almos negligible e e sibili y caused
by he i e e sible o ma ion o LiFeO
2
, while he la e showed
wo se kine ics and e e sibili y due o he o ma ion o po assium
be a e i e. Fo hese easons, he MnFe
2
O
4
eNa
2
CO
3
was kep as
Fig. 2. (a) Mass p ofiles o he MnFe
2
O
4
eNa
2
CO
3
mix u e du ing 10 deca bona ion-ca bona ion cycles pe o med be ween 800 C and 400 C. (b) Mass change ela ed o
deca bona ion and ca bona ion s eps du ing he cycles. (ce ) SEM images o he MnFe
2
O
4
eNa
2
CO
3
mix u e a e cycle 3. The images we e acqui ed a e he hi d deca bona ion
(ced) and (ee ) a e he hi d ca bona ion s eps, espec i ely. (g) Schema ic ep esen a ion o he sin e ing and phase coalescence p ocesses ha lead o he dec ease o
e e sibili y du ing he fi s cycles. SEM, scanning elec on mic oscopy.
F. To e, T.A. Sanchez, S. Doppiu e al. Ma e ials Today Ene gy 29 (2022) 101094
6
he s a ing e e ence sys em o in es iga e he a omic-subs i u ed
spinel e i es ha a e p esen ed in he ollowing sec ion.
3.3. E ec o Mn subs i u ion
In his sec ion, he e ec o pa ial subs i u ion o Mn o Zn, Ni,
and Ca on he deca bona ion-ca bona ion eac ion is p esen ed.
Samples a 5 and 10 a .% o Zn Ni, and Ca we e cycled 10 imes. Da a
a e epo ed in Fig. 4 by aking he MnFe
2
O
4
eNa
2
CO
3
mix u es as
he e e ence sys em. In he fi s cycle, he h ee 5 a .%-doped
samples show sligh ly be e deca bona ion kine ics han he
undoped sys em, which can be app ecia ed by he small le shi o
he mass p ofiles in he ime scale (Fig. S3a). Howe e , such di -
e ence is no u he app ecia ed in he fi h and nin h cycles
(Fig. S3b and c), as he ou deca bona ion p ofiles almos o e lap.
Same hing o he ca bona ion kine ics; he small di e ences be-
ween he ou samples ha a e obse ed du ing he fi s cycle,
p og essi ely anish as he numbe o cycles inc eases.
Majo di e ences can be app ecia ed conce ning he amoun o
deso bed CO
2
, as can be app ecia ed in Fig. 4a. In he fi s cycle, he
undoped MnFe
2
O
4
shows he highes mass loss (9.76 w %), ol-
lowed by he 5% Ni, 5% Zn, and 5% Ca e i es, which exhibi a mass
loss o 8.8, 8.15, and 6.8 w %, espec i ely. Howe e , he undoped
MnFe
2
O
4
p og essi elyloses capaci y while he h ee doped e i es
seem o su e less e e sibili y loss, wi h he Zn-doped sample
e en showing a sligh inc ease in pe o mance. A e 10 cycles, he
undoped MnFe
2
O
4
and he 5% Ni samples show a simila beha io ,
as hey lose 8.14 w % and 8.2 w %, espec i ely. The Ca-doped
sample deso bs only 6.5 w %, while he Zn-doped leads o he
highes mass loss (8.6 w %).
Fig. 3. Mass p ofiles o he MnFe
2
O
4
eNa
2
CO
3
, MnFe
2
O
4
eLi
2
CO
3
, and MnFe
2
O
4
eK
2
CO
3
mix u es du ing (a) he fi s deca bona ion s ep and (b) du ing 3 deca bona ion-ca bona ion
cycles. (c) XRD pa e ns collec ed o he h ee mix u es be o e and a e 3 cycles. XRD, X- ay di ac ion.
Fig. 4. The e ec o 5 and 10 a .% Ca, Ni, and Zn a omic subs i u ion on he MnFe
2
O
4
eNa
2
CO
3
mix u e du ing 10 deca bona ion-ca bona ion cycles. (a) The measu ed CO
2
elease
du ing 10 cycles o he 5 a .% Ca, Ni, and Zn-doped e i es, and (b) o he 10 a .% Ca, Ni, and Zn-doped e i es. (c) XRD pa e ns o di e en Mn
1-x
A
x
Fe
2
O
4
eNa
2
CO
3
mix u es a e
10 deca bona ion-ca bona ion cycles. XRD, X- ay di ac ion.
F. To e, T.A. Sanchez, S. Doppiu e al. Ma e ials Today Ene gy 29 (2022) 101094
7
Da a o he mix u es wi h 10 a .% o Zn, Ni, and Ca a e epo ed
in Fig. 4b and Fig. S3d. O e all, he inc ease in a omic subs i u ion
om 5% o 10% does no lead o any significan change. A e 9
cycles, he h ee subs i u ed e i es exhibi a sligh wo sening in
he mass loss a e in compa ison o pu e MnFe
2
O
4
(Fig. S3d), while
he ca bona ion eac ion p oceeds as e . Among he doped e i es,
no significan di e ences can be app ecia ed in e ms o kine ic
pe o mances. On he o he hand, he dopan elemen a ec s he
e e sibili y as he 10% Zn and 10% Ni e i es show a simila weigh
loss du ing all he cycles, while he subs i u ion o Ca esul s in a
lowe CO
2
elease. All he samples show a loss o e e sibili y ha is
mo e e iden du ing he fi s cycles. A e 10 cycles, pu e MnFe
2
O
4
shows he highes deso p ion, ollowed by 10% Zn, 10% Ni, and 10%
Ca samples ha lose 7.9 w %, 7.8 w %, and 7 w %, espec i ely.
Acco ding o he XRD analysis, he pa ial Mn subs i u ion o Ni
and Zn did no induce significan changes in he phase composi ion
o he mix u es (Fig. 4c). No app eciable amoun s o seconda y
phases we e de ec ed sugges ing ha he ini ial fine dispe sion o
he dopan s was e ained e en upon p olonged cycling. On he
o he hand, he XRD pa e ns o bo h he 5 a .% and 10 a .% Ca-
doped e i es show he o ma ion o Ca
2
Fe
2
O
5
as a pa asi e
phase, which explains he dec ease in pe o mances obse ed
du ing he cycles.
The XRD pa e ns o he di e en Mn
1-x
A
x
Fe
2
O
4
eNa
2
CO
3
mix-
u es we e efined o ge u he insigh . The a e age la ice
pa ame e o he wo main phases, i.e. he cubic MnFe
2
O
4
and he
igonal NaFeO
2
, a e epo ed in Fig. S3e and , espec i ely. O e all,
a gene al dec ease o he wo la ice pa ame e s is obse ed as he
amoun o dopan inc eases. While his beha io is less e iden o
Zn e i es, i becomes way mo e significan o he samples doped
wi h Ni and Ca. Da a ob ained o Zn and Ni samples a e in line wi h
wha was expec ed, as a simila end was obse ed o he la ice
pa ame e o he as-syn hesized Ni and Zn e i es (Sec ion S1
Fig. S1b). On he con a y, da a a e mo e di ficul o in e p e in
he case o Ca-doped e i es. Indeed, he as-syn hesized
CaeMnFe
2
O
4
samples showed a significan la ice expansion due
o he la ge Ca
2þ
ions in compa ison o Mn
2þ
. Fo his eason, he
la ice sh inkage de ec ed a e 10 cycles would no be expec ed. A
possible in e p e a ion can be ound in he p ecipi a ion o he
Ca
2
Fe
2
O
5
phase, which may lead o he o ma ion o a high con-
cen a ion o acancies in he NaFeO
2
phase. In gene al, a
s aigh o wa d connec ion be ween he la ice olume and he
deca bona ion-ca bona ion pe o mances can be excluded o he
h ee sys ems in es iga ed.
3.4. Hyd ogen p oduc ion
Among he composi ion in es iga ed in he p e ious sec ion, he
spinel wi h 5% o Zn showed he bes e e sibili y du ing he
deca bona ion-ca bona ion cycles. This composi ion was hen
selec ed o WS expe imen s. In he ollowing, he H
2
p oduc ion o
he Zn
0.05
Mn
0.95
Fe
2
O
4
eNa
2
CO
3
sys em is p esen ed, while using
he MnFe
2
O
4
eNa
2
CO
3
mix u e as a e e ence (Fig. 5).
The samples we e ini ially es ed in dynamic condi ions, by
modi ying he empe a u e p og am used in he deca bona ion-
ca bona ion cycles (see Sec ion 2.2). The wo mix u es we e
cycled en imes o in es iga e he e e sibili y o he H
2
p o-
duc ion eac ion. The fi s WS s ep is epo ed in de ail in
Fig. 5a, whe e he mass p ofiles a e shown oge he wi h he H
2
e olu ion. The gas flow composi ion co esponding o each s ep
o he expe imen is also epo ed o acili a e da a in e p e a-
ion. Du ing he fi s 100 min o he expe imen , no wa e apo
was in oduced. As p e iously ca ied ou o he 1:1 mix u es,
he samples we e hea ed o 800
C and kep unde iso he mal
condi ions o 30 min.
As can be app ecia ed in Fig. 5a, an ini ial small mass loss o 0.45
and 0.75 w % is obse ed o undoped and Zn-doped mix u es,
espec i ely; his is likely due o he deso p ion o some esidual
mois u e om Na
2
CO
3
powde s, which a e highly hyg oscopic. As
expec ed, bo h mix u es s a eleasing CO
2
a a ound 550
C; he
mass loss a e fi s inc eases and hen app oaches a pla eau a e
a ound 30 min a 800
C. In his ega d, he sample wi h 5% o Zn
shows a be e kine ic han he undoped MnFe
2
O
4
,aswellasa
highe CO
2
deso p ion (12. w .% s 10 w %). Such alues a e in line
wi h he heo e ical mass loss expec ed o he elease o 2/3 o he
o al CO
2
, which acco ding o eac ion 1, co esponds o a weigh
loss o 11.3 w %. This confi ms he mechanism p oposed by Va sano
e al. and esumed in he in oduc ion ( eac ions 3 and 4) [23]. As
soon as wa e apo is in oduced, he masses o he wo samples
s a o dec ease again and H
2
is suddenly de ec ed. The amoun o
H
2
inc eases and eaches a maximum o 22.3 and 12.6
m
mol g
1
s
1
o pu e and 5% Zn MnFe
2
O
4
, espec i ely. E en i a a lowe a e, H
2
p oduc ion con inues o bo h samples as long as wa e apo is
supplied. In he mean ime, he masses o he wo samples keep
dec easing and s abilize a a ound 85% o he ini ial alue. An
o e all mass loss o 14.4 w % and 15.2 w % is de ec ed o he
undoped and Zn-doped mix u es, espec i ely. Acco ding o he
s oichiome y o eac ion 1, heo e ical mass losses o 14.9 and
15 w % a e expec ed o he undoped and Zn-doped mix u es,
espec i ely. These alues a e gi en by he sum o wo di e en
con ibu ions. Fi s , he CO
2
elease due o he decomposi ion o
Na
2
CO
3
, which co esponds o a 16.9 w % mass loss. Second, he O
2
up ake om he eac ion wi h wa e s eam o o m he NaMn
1/3
Fe
2/3
O
2
, which co esponds o a 2 w % mass gain. The second
con ibu ion is linked o he oxida ion o Mn
2þ
o Mn
3þ
and, in u n,
o he H
2
p oduc ion. As Zn does no con ibu e o he edox p o-
cess, a 5%-lowe mass change is expec ed o he 5% Zn
spineldi.e. þ1.9 w %. Conside ing he addi ional mass change ha
is due o he ini ial mois u e deso p ion, a maximum mass loss o
15.35 and 15.78 w % a e easonably expec ed o he undoped and
he Zn-doped mix u es; compa ing he expe imen al da a wi h
hese alues, con e sions o 93.8% and 96.3%, espec i ely, we e
ob ained.
The da a ob ained om he he mobalance we e compa ed wi h
hose ob ained om he H
2
mic osenso . In pa icula , he in e-
g a ion o he H
2
p oduc ion o e ime p o ided H
2
yields o
1.14 mmol/g o pu e MnFe
2
O
4
and 0.66 mmol/g o 5% Zn MnFe
2
O
4
.
As he heo e ical yield o he MnFe
2
O
4
eNa
2
CO
3
mix u e is
1.28 mmol/g, a ela i e H
2
yield o z89% is ob ained, which ag ees
wi h he alue ob ained om he mog a ime ic da a. On he o he
hand, he Zn-doped mix u e ma ks an H
2
yield o 54% only, wi h i s
heo e ical H
2
yield being 1.21 mmol/g. Such alue is significan ly
lowe han he one calcula ed based on he mog a ime y and i is
coun e in ui i e a a fi s glance. In ac , as he Zn-doped spinel
showed a highe mass loss han he undoped, one would expec a
highe H
2
yield o he o me . Such esul s can be in e p e ed by
ecalling he di e en con ibu ions o he mass change, meaning
CO
2
elease and O
2
up ake.
The o e all di e ence be ween he expe imen ally obse ed
mass loss o he Zn-doped and ha o he undoped mix u e is
0.8 w %. This alue educes o 0.5 w % when conside ing he con-
ibu ions o he ini ial mois u e deso p ion. XRD analysis was
pe o med a e he fi s WS s ep (Fig. S4a). Fo bo h mix u es,
NaMn
1/3
Fe
2/3
O
2
was iden ified as he main phase, ollowed by
mino amoun s o un eac ed Na
2
CO
3
and MnFe
2
O
4
. I is hen
easonable o assume ha he CO
2
elease o he 5% Zn-doped
mix u e ook place wi h no significan changes compa ed o he
undoped. The di e ence in e ms o mass loss be ween he wo
mix u es is hen ela ed only o he O
2
up ake, which is also sup-
po ed by he high di e ence in e ms o p oduced H
2
. A e a ew
F. To e, T.A. Sanchez, S. Doppiu e al. Ma e ials Today Ene gy 29 (2022) 101094
8
simple p opo ions, his leads o a ela i e H
2
yield o 63%, which is
in be e ag eemen wi h he alue ob ained by he di ec in e-
g a ion o he H
2
signal.
Du ing he subsequen cycling in dynamic condi ions, he Zn-
doped mix u e showed a be e e e sibili y in e ms o mass
change (Fig. S4b) ha also co esponded o a significan ly lowe
amoun o undesi ed seconda y phases such as he non-
s oichiome ic Na
x
Mn
3
O
7
(Fig. S4c). Howe e , H
2
was no de ec-
ed du ing he ollowing 9 cycles o bo h samples. Despi e his, he
p esence o significan amoun s o MnFe
2
O
4
and Na
2
CO
3
in he
cycled mix u es sugges s ha e en i wi h a lowe yield, H
2
was
p oduced also du ing he emaining cycles. A possible explana ion
is ha he incomple e egene a ion o he eac an s lowe ed he H
2
p oduc ion so ha he H
2
concen a ion was lowe han he
de ec ion limi o he mic osenso used. A mo e de ailed discussion
is p o ided in Sec ion 4.1 o he suppo ing In o ma ion file.
Despi e he e ec o Zn-doping on he H
2
p oduc ion could no
be di ec ly e alua ed in he abo e-men ioned expe imen al con-
di ions, bo h XRD and he mog a ime y sugges ed a posi i e e ec
o Zn on he e e sibili y. To shed ligh on his, bo h he undoped
MnFe
2
O
4
eNa
2
CO
3
and he 5% Zn-doped MnFe
2
O
4
eNa
2
CO
3
mix-
u es we e subsequen ly es ed in iso he mal condi ions a 750
C
o a o al o 5 cycles (Sec ion 2.2.2 o de ails). The H
2
p oduc ion o
he wo mix u es is epo ed in Table 2, and he da a a e plo ed in
Fig. 5b. As seen be o e o he expe imen s a 800
C, he undoped
MnFe
2
O
4
eNa
2
CO
3
mix u e p oduces mo e H
2
du ing he fi s cycle
(86% yield). Howe e , i s H
2
p oduc ion d as ically dec eases by wo
o de s o magni ude in he second cycle and keeps dec easing
du ing he ollowing h ee. E en ually, almos no H
2
is de ec ed in
he fi h cycle. On he o he hand, he Zn-doped mix u e p oduces
less H
2
du ing he fi s cycle (54% yield) bu i s p oduc ion s ays one
o de o magni ude highe han he undoped mix u e du ing he
ollowing 4 cycles. Mo eo e , exemp om a fi s d op a e he fi s
cycle, no dec easing endency can be obse ed in he ollowing
cycles, and he Zn-doped mix u e shows a s able H
2
p oduc ion o
z0.2 mmol/g om cycle 2 o cycle 5. High e e sibili y is obse ed
also in e ms o kine ic, as he H
2
p oduc ion p ofiles om cycles 2
o 5 almos pe ec ly o e lap (Fig. S5). This is also in line wi h he
high e e sibili y obse ed in e ms o deso bed/cap u ed CO
2
(Table S1). As a esul o he wo e y di e en beha io s, he di -
e ence in e ms o H
2
p oduc ion pe o mances be ween he wo
mix u es becomes mo e e iden du ing he cycles. In he second
cycle, he Zn-doped mix u e p oduces 4.6 imes mo e H
2
han he
undoped one, while in he fi h cycle, i p oduces 23 imes mo e.
This comple ely di e en end also a ec s he cumula i e H
2
p oduc ion, which is highe o he Zn-doped mix u e (1.60 s.
1.24 mmol/g).
4. Conclusions
In he p esen wo k, we s udied he e ec o a omic subs i u ion
on he sodium manganese e i e he mochemical cycle o H
2
p oduc ion. The deca bona ion/ca bona ion o he
MnFe
2
O
4
eNa
2
CO
3
mix u e was in es iga ed as a s a ing e e ence,
while bo h Na and Mn subs i u ion we e subsequen ly conside ed.
Na subs i u ion o K led o a dec ease in pe o mance ha was
a ibu ed o he o ma ion o po assium be a e i e as an unde-
si ed phase. On he o he hand, he use o Li dec eased he onse
deca bona ion empe a u e by abou 100
C, bu negligible
e e sibili y was obse ed unde he explo ed expe imen al con-
di ions. Despi e his, he subs i u ion o Na o Li is p omising o
dec easing he ope a ing empe a u e o he he mochemical cycle
and dese es u he in es iga ion.
Mn pa ial subs i u ion o Ca, Ni, and Zn imp o ed he a e o
he ca bona ion eac ion. The Mn
0.95
Zn
0.05
Fe
2
O
4
eNa
2
CO
3
mix u e
showed he bes e e sibili y and was es ed o H
2
p oduc ion,
while he undoped mix u e was used as he e e ence. Du ing he
fi s cycle, maximum ins an aneous H
2
p oduc ion a es o 22.3 and
12.6
m
mol g
1
s
1
we e obse ed o he undoped and he Zn-
doped samples. A e 1.5 h a 800
C, he MnFe
2
O
4
eNa
2
CO
3
mix u e eached 89% o he heo e ical yield, wi h an o e all H
2
p oduc ion o 1.14 mmol/g. On he o he hand, only 0.66 mmol H
2
/g
(54% con e sion) we e p oduced du ing he same ime by he Zn-
doped mix u e. In bo h cases, cycling in dynamic condi ions
(800e400
C) did no ensu e a comple e egene a ion o he
Fig. 5. (a) Mass p ofiles and H
2
e olu ion o he MnFe
2
O
4
eNa
2
CO
3
and 5% Zn MnFe
2
O
4
eNa
2
CO
3
mix u es du ing he fi s WS s ep a 800 C. (b) H
2
p oduc ion (mmol H
2
/g) o he
undoped MnFe
2
O
4
eNa
2
CO
3
and he 5% Zn-doped mix u es du ing 5 cycles a 750 C.
Table 2
H
2
p oduc ion (mmol H
2
/g) o he undoped MnFe
2
O
4
eNa
2
CO
3
and he 5% Zn-doped
mix u es du ing 5 cycles a 750
C.
Mix u e Hyd ogen p oduc ion (mmol H
2
/g)
Cycle 1 Cycle 2 Cycle 3 Cycle 4 Cycle 5 To al
MnFe
2
O
4
1.10 0.05 0.04 0.03 0.01 1.24
5% Zn 0.69 0.23 0.22 0.22 0.23 1.60
F. To e, T.A. Sanchez, S. Doppiu e al. Ma e ials Today Ene gy 29 (2022) 101094
9
