ethanol 5ad aafsdfastfwqfas sacaxvava axcvvvvvvvvvvvvvvvvasddddddddddd

8/10/2019 ethanol 5ad aafsdfastfwqfas sacaxvava axcvvvvvvvvvvvvvvvvasddddddddddd

http://slidepdf.com/reader/full/ethanol-5ad-aafsdfastfwqfas-sacaxvava-axcvvvvvvvvvvvvvvvvasddddddddddd 1/7

Hydrogen from ethanol by steam iron process inxed bed reactor

E. Hormilleja, P. Dura´ n, J. Plou, J. Herguido, J.A. Pena*Catalysis, Molecular Separations and Reactor Engineering Group (CREG), Aragon Institute of Engineering Research(I3A), Universidad Zaragoza, Mariano Esquillor s/n, Ed. Iþ D, 50018 Zaragoza, Spain

a r t i c l e i n f o

Article history:Received 27 September 2013Received in revised form18 December 2013Accepted 1 January 2014Available online 1 February 2014

Keywords:Bio-ethanolHydrogenSteam-IronRedox

Chemical looping

a b s t r a c t

This research is devoted to the use of ethanol (i.e. bio-ethanol) in the combined productionand purication of hydrogen by redox processes. The process has been studied in a singlelab scale xed bed reactor. Iron oxides, apart from their remarked redox behavior, exert animportant catalytic role allowing the complete decomposition of ethanol at temperaturesin the range from 625 to 750 C. The resulting gas stream (mainly H 2 and CO) reduces thesolid to metallic iron. During a subsequent oxidation with steam, the solid can be regen-erated to magnetite producing high purity hydrogen (suitable to be used in PEM fuel cells).Even though small amounts of coke are deposited during the reduction step, this is barelygasied by steam during the oxidation step (detection of CO x in concentrations lower than1 ppm). Inuence of parameters like temperature, ethanol partial pressure and alternatecycles’ effect has been studied in order to maximize the production of pure hydrogen.Copyright ª 2014, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights

reserved.

1. Introduction

The present study is focused in the joint production and pu-rication of hydrogen from renewable ethanol. Ethanol is agreat option which can be converted into hydrogen due to itsrelative low-toxicity (lesser than that of methanol), easy-generation from renewable sources, and low productioncosts [1].

Nowadays, most ethanol is produced by fermentationprocesses using raw materials like sugar cane, cereal grainand other materials with high starch content [2]. This way,biomass has produced up to 85.2 million cubic meters per yearofbio-ethanol in 2012 all over theworld [3], being currently themost used biofuel in transportation [4]. To refer to this

commodity the term “ bio-ethanol” has been adopted not onlyto specify its biological origin but also or primarily as a mar-keting resource. During the last decade, ethanol has attractedconsiderable attention given that its research has achievedinteresting improvements through the so called sustainable2nd generation bio-fuels (ethanol obtained from lignocellulosicmaterials) [5]. This research has been the consequence of

trying to avoid the drawbacks derived from the use of agro-food assets in the production of fuels, and to confer addedvalue to crops and lands of low productivity [6].

The main process to produce hydrogen from ethanol in-volves steam reforming [7]. Nevertheless, an original methodhas beenproposed in thenear past that consists in its thermalcatalytic decomposition [1]. This method takes advantage notonly of the hydrogen production itself, but also of the

* Corresponding author . Tel.: þ 34 976 762390; fax: þ 34 976 762043.E-mail address: [email protected] (J.A. Pen ˜ a).

Available online at www.sciencedirect.com

ScienceDirect

journal homepage: www. e l s ev i er. com/ lo ca t e /he

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 9 ( 2 0 1 4 ) 5 2 6 7 e 5 2 7 3

0360-3199/$ e see front matter Copyright ª 2014, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.ijhydene.2014.01.002

mailto:[email protected]

http://www.sciencedirect.com/science/journal/03603199

http://www.elsevier.com/locate/he

http://dx.doi.org/10.1016/j.ijhydene.2014.01.002






http://www.elsevier.com/locate/he

http://www.sciencedirect.com/science/journal/03603199

http://crossmark.crossref.org/dialog/?doi=10.1016/j.ijhydene.2014.01.002&domain=pdf

mailto:[email protected]



coproduction of carbon nanostructured materials (carbonnanobers e CNF and nanotubes e CNT). Although, it hasbeen studied by several research groups [1,8], the drawback of lowhydrogen purityby signicantpresence of CO andCO 2 stillremains.

Thegoalsought in this work tries to cope with this problemthrough the simultaneous production and purication of

hydrogen from ethanol by the steam iron process (SIP) [9,10],in a xed bed reactor. This process is based on alternate cyclesof solid reduction and oxidation. During reduction steps, ahighly reductive stream (ethanol) reacts with iron oxides toproduce a reducedform of the solid andvariable compositionsof the exhaust gases. On a later step, the previously reducediron is re-oxidized with steam to form magnetite and a highpurity hydrogen stream [4]. This process is closely related alsoto the chemical looping reduction, a variant of the chemicallooping combustion using steam to re-oxidized the solid andreleasing hydrogen [11].

1.1. Reaction network

Based on literature [1,8,12], the reactions that probably betterdescribe the process of ethanol decomposition along the ex-periments are the following:

Thermal decomposition:

CH3CH2OH# COþ H2 þ CH4 (r.1)

Ethanol dehydration :

CH3CH2OH# CH2 ¼ CH2 þ H2O (r.2)

Besides, these products could generate the following reactions:

CH4# CðsÞ þ 2H2 (r.3)

CH2 ¼ CH2# 2CðsÞ þ 2H2 (r.4)

The sum of reactions (r.1) plus (r.3), and (r.2) plus (r.4),would conform the ethanol decomposition reactions (r.5) and(r.6):

CH3CH2OH# CðsÞ þ COþ 3H2 (r.5)

CH3CH2OH# 2CðsÞ þ 2H2 þ H2O (r.6)

On theother hand ethylene from (r.2) could be hydrogenatedas shown in (r.7):

CH2 ¼ CH2 þ H2# CH3 CH3 (r.7)

Methanol could come from the carbonylation of ethanol toethyl formate, followed by catalytic hydrogenolysis of the ethylformate to methanol and ethanol [12] resulting in:

2H2 þ CO# CH3OH (r.8)

The presence of CO 2 could result also from steam reformingof ethanol if water can be present in the reacting atmosphere:

CH3CH2OH þ 3H2O# 6H2 þ 2CO2 (r.9)

Ethanol decomposition, as described by (r.1), (r.5) and (r.6)generates a gas stream rich in hydrogen and carbon monox-ide with a considerable reductive potential that could allow

the reduction of the solid starting from hematite up tometallic iron. This would constitute the rst step of the abovementioned SIP.

At the experimental conditions used throughout this work(600e 750 C and 1 bar), iron oxide reduction by hydrogen ex-hibits two stages [9,13]: The rst one consists in the reductionfrom hematite (Fe 2O3) to magnetite (Fe 3O4) (r.10). The second

is the reduction from magnetite to metallic iron (r.11). Similarbehavior can be attained when CO is the reducer gas (r.12) and(r.13). Although reduction of intermediate wustite was ex-pected at temperature higher than 570 C [14] according tothermodynamic solid phase diagrams [15], it was not empiri-cally evidenced.

3Fe2O3ðsÞ þ H2 # H2O þ 2Fe3O4ðsÞ (r.10)

Fe3O4ðsÞ þ 4H2 # 4H2O þ 3FeðsÞ (r.11)

3Fe2O3ðsÞ þ CO# CO2 þ 2Fe3O4ðsÞ (r.12)

Fe3O4ðsÞ þ 4CO# 4CO2 þ 3FeðsÞ (r.13)During the second step of steam iron process, the reduced

iron is oxidized by steam producing pure hydrogen. Theformer solid is regenerated only to intermediate magnetitedue to thermodynamic restrictions at experimental temper-aturesand partial pressures (reverse of r.11 , now on labeled asr.11*).

It is important to note that also reactions between gaseousreactants and products can take place within the reactor.These could produce reactions such as methanation (r.14),Water Gas Shift (r.15) or Boudouard reaction (r.16).

3H2 þ CO# H2O þ CH4 (r.14)

H2O þ CO# H2 þ CO2 (r.15)

2CO# CO2 þ CðsÞ (r.16)

Finally, deposited coke during the reduction step could begasied during the subsequent oxidation step, as shown in(r.17).

CðsÞ þ H2O# COþ H2 (r.17)

2. Experimental

2.1. Experimental setup

The experimental system consisted of a cylindrical xed bedreactor made out of quartz ( Øi ¼ 13mm, L¼ 420mm).The solidbed is constituted by 2.125 g: 75%w “ triple” oxide (Fe2O3, Al2O3

and CeO 2). Its length is roughly 2 cm. The usage of an additi-vated solid (cerium and aluminum oxides),Øparticle ¼ 160e 200 mm, has been proposed in order to improvethe stability to sintering of the solid [16]. The other 25%w isSiO2 of same diameter acting as inert. Silica is added to avoidagglomeration of the reacting solid upon reoxidation and todecrease preferential pathways of the gas stream in its cross-sectional distribution. No temperature proles greater than








2 C were detected along the bed in any of the steps (reductionor oxidation) ensuring an isothermal behavior.The solid bed issettled over a quartz porous plate ( Ømesh ¼ 90 mm).

In both reduction and oxidation experiments, the reactionsystem was operated at atmospheric pressure. Total gas ow(250 NmL mine 1, large enough to avoid external diffusionalcontrol) was fed from top of the reactor (WHSV ¼ 2.0 h 1;GHSV ¼ 5.7 103 h 1). Ethanol was used in the reductionsteps (partial pressure of 0.05, 0.1 and 0.15 bar;_nCH3CH2OH

in ¼ 1:12 mmol min e 1 for 0.1 bar) diluted in He, whilewater was fed in the oxidation steps (partial pressure of 0.25 bar; _nH2O

in ¼ 2:79 mmol min e 1) also diluted in He.Ar (partial

pressure of 0.05 bar) was used as internal standard. Ethanolandwater were dosed by respective HPLC pumps Shimadzu LC-20AT. After that, the liquids were vaporized and fed to thereactor. Inert gases (He and Ar) were added through mass owmeters Brooks 5850 and Alicat Scientic.

The reactor was contained inside an electric oven. Tem-perature interval for reduction experiments was set from625 C to 750 C, while the oxidation step was always per-formed at 500 C. The experiments at different reductiontemperatures were carried out along three successive reduc-tion and oxidation cycles, while the analysis of the inuenceof ethanol partial pressure was conducted through one redoxcycle.

Exhaust gases (including steam and ethanol) wereanalyzed using a gas chromatograph (TCD) Agilent 7890A.

2.2. Solid preparation

The reactive solid (“ triple”) was a mixture of 98%w Fe 2O3, withtwo additives: 1.75%w Al 2O3 and 0.25%w CeO2. Alumina de-creases sintering effectsover the iron oxide,which results in ahigher efciency and stability [16]. Ceria promotes and in-creases redox reaction rates. This formula was previouslytested and validated as optimal for high performance inhydrogen purication from methanol decomposition [17].

The synthesis of the triple oxide was carried out by amethod based in the precipitation of citrates from metallicnitrates. Obtained gelwasdried at 60 C (12 h), then calcined to800 C (b ¼ 5 C min e 1) and nally sieved to 160 e 200 mm par-ticle diameter. More details about the synthesis route andproportions of reagents can be found in literature [17,18].

3. Results

3.1. Thermal decomposition of ethanol

Thermal decomposition of ethanol was checked out inabsence of solid at 700 C. Stable conversion of ca. 46% was

achieved along the time (for more than 3 h of time on stream).Methanol and hydrogen were the majority products, followedby a smaller generation of CO, CH 4 and H 2O. In addition, aminimumquantity of ethylene and ethane (selectivities in theorder of 1%) were observed. CO 2 presence was not gured out,but a little deposition of coke over the porous plate wasvisually detected after the experiment.

Table 1 includes the major species for a blank experiment(thermal decomposition without solid), as well as the resultsobtained using the same processing conditions with “triple”oxide during the reduction step. Ethanol molar feed was al-ways 1.12 mmol min 1 for a partial pressure of 0.1 bar.

3.2. Solid reduction with ethanol

Fig. 1 shows the composition of the exhaust gases during therst reduction step at 675 C. Since no ethanol was detected atthe exit, it might be concluded that the solid plays an activerole, probably through decomposition reactions (r.5) and (r.6),and by consumption of lattice oxygen of Fe 2O3. During a rststage (labeled as Stage “A” in Fig. 1) and for 5 min, reduction of hematite to magnetite follows reactions (r.10) and (r.12)through emerging H 2 and CO. This point will be documentedlater on.

During the following stage ( Stage “B” in Fig. 1), solid isprogressively being reduced from magnetite to metallic ironas described by reactions (r.11) and (r.13). Reaction rates arequite lower than in the previous stage, showing a slight in-crease in selectivity towards H 2 and CO along time. Bothstages extend up to the rst 25 min (end of solid reduction).

Table 1 e Molar ow ratios of different species at the exit vs. ethanol feed at the entrance of the reactor fora blank (without solid) and with a sample of “triple” oxide in the reduction step.[Adim] _nout

H2O_nin

CH3CH2OH

_noutCH4

_ninCH3CH2OH

_noutCO

_ninCH3CH2OH

_noutCO2

_ninCH3CH2OH

_noutCH3OH

_ninCH3CH2OH

_noutH2

_ninCH3CH2OH

_noutC2H4

_ninCH3CH2OH

_noutC2H6

_ninCH3CH2OH

Blank 0.133 0.180 0.184 0 0.364 0.327 0.087 0.005Triple oxide 0.119 0.107 0.734 0.784 0 2.899 0 0.004

0 5 10 15 20 25 30 35 400.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

F CO máx .

Stage "C" Stage "B"

H2

H2O CO2

CO CH4

F H2 máx .

M o

l a r

f l o w

( m m o

l / m i n )

Time (min)

Stage "A"

0.0

0.2

0.4

0.6

0.8

1.0

Δ P

O v e r p r e s s u r e

( b a r - g

)

Fig. 1 e Product distribution and overpressure prole alongtime for the reduction of fresh iron oxide with ethanol(1.12 mmol min e 1 , 0.1 bar) at 675 C.

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 9 ( 2 0 1 4 ) 5 2 6 7 e 5 2 7 3 5269







From this time on, molar ows show a stable behavior withhigh values for H 2 and CO, very close to the stoichiometricvalues (3.36 mmol min 1 and 1.12 mmol min 1, respectively,at the operating conditions) provided by the decomposition of ethanol according to reaction (r.5) (Stage “C” in Fig. 1). Conse-quently low selectivities (ca. 3%) towards oxidized species(H2O and CO2) were observed in this same period.

The deviation in the elemental balance (e.g. relative dif-ference between output molar ow and input molar ow foreach component) for C, H and O performed at the exit of thereactor on gas phase during a reduction experiment ( Fig. 2),shows that a good balance closure has been achieved forhydrogen. However, high excess of oxygen is obtained during the rst reduction minutes due to the oxygen contributioncoming from the solid.Carbonshows an important decit (e.g.carbon accumulation within the reactor) from the verybeginning, stabilizing at constant closure deviation from10 min on. This indicates that carbon deposition proceeds atan almostconstant rate over solid particle surface. Presence of amorphous as well as structured nanotubes has been evi-denced by TEM after each cycle (not shown), being of similarcharacteristics as those found in analogous studies using methanol as test reagent [17]. Although iron carbides weresought by XRD of coked samples after reduction, they werenever detected.

This coke deposition causes a slight overpressure increasein the bed (dashed curve in Fig. 1) during the rst minutes of reaction, stabilizing later on with inconsequential values(close to 0.3 bar-g at experimental conditionscorresponding toFig. 1). Only for longer times on stream , and always within Stage“C” would be found a deviation from this behavior,showing anabrupt increase in pressure drop (not presented here) thatcould jeopardize the operation of the reactor. Nevertheless, itshould be noted that actually Stage “C” is not necessary for thesteam iron process performance; therefore the reduction stepshould be interrupted when reaching this Stage “C” in order toavoid overpressure problems.

3.3. Solid reoxidation with steam and H 2 generation

After the reduction step in which iron oxide is transformedinto metallic iron, the next step consists of an oxidation withsteam [19], according to reaction (r.11*). Fig. 3 shows thetemporal evolution of steam and products (H 2 and COx ) at500 C after the reduction step shown in Fig. 1 (675 C). As timegoes by, H 2 production decreases while steam rises in thesame proportion. At the end of the experiment, the maximumtheoretical value of hydrogen production is almost reached(26.20 mmol H2 vs 26.24 mmol H 2). In addition, only a minimalamount of CO and CO 2 was detected in exhaust gases by GC(concentrations always lower than 1 ppm), which maintainsby far the purity standards required for the application of hydrogen in PEMFC, and reects the inert character of thecoke produced in the reduction step. Carbon gasication byreaction (r.17) results to be minimum.

3.4. Inuence of partial pressure of ethanol

Fig. 4 shows the inuence of ethanol partial pressure (0.05,0.10 and 0.15 bar) over hydrogen and water molar ows during the reduction of fresh solid along time ( Tred ¼ 675 C). Thethree characteristic stages of fresh solid reduction ( Stages “A”,“B” and “C” in Fig. 1) can be distinguished each one with theirown values but showing the same trend. As it could be fore-seeable, the higher the ethanol partial pressure, the higher thereduction rate. The stable behavior correspondent to Stage “C”in Fig. 1 also depends on the reactant partial pressure, taking longer to reach it for those experiments with lower ethanolpartial pressures at the entrance.

3.5. Inuence of reduction temperature

The complete reduction of fresh solid is reached in a shortertime by increasing its temperature. This effect is shown inFig. 5, where it is represented the effect of temperature on thetemporal evolution of water molar ow at the exit of thereactor during the reduction step of fresh triple oxide with

0 5 10 15 20 25 30 35 40-75

-50

-25

0

25

50

75

100

210

220

M a s s

b a

l a n c e c

l o s u r e

( % )

Time (min)

C H O

Fig. 2 e Elemental mass balance closures (output over input molar ows) for reduction of fresh iron oxide with ethanol(0.1 bar) at 675 C.

0 10 20 30 40 50 60 70 80 90

0.0

0.5

1.0

1.5

2.0

2.5

3.0

< 1 ppm CO x

H2O

H2

CO2

CO

M o

l a r

f l o w

( m m o l / m

i n )

Time (min)

F H2O,0

Fig. 3 e Reactive and products distribution during the rst oxidation step at 500 C (previous reduction step at 675 C).








ethanol. At the same time, the graph shows in dashed linesthe thermodynamic theoretical values corresponding toethanol catalytic decompositionequilibrium corresponding toStage “C” in Fig. 1. This way, the proximity between experi-mental curves and theoretical equilibrium values is veried.

Nevertheless, reaction rates during the oxidation stepbecome slower by increasing the temperature of the previousreduction step. This fact is a consequence of sintering, whichresults in a decrease of the solid reactivity. This is shown inFig. 6 where it is represented the temporal evolution of hydrogen molar ow during the oxidation step (500 C).

Furthermore, it has been found that all curves in Fig. 6 presenta coincident molar ow value at approximately 10 min of re-action. This time ( t*) corresponds to the minimum theoreticaltime to attain the complete oxidation of the solid, whichwould be achieved if reaction rate were maximum.

3.6. Inuence of redox cycles

Experimental series consisting of three consecutive redoxcycles were conducted at different temperatures for re-ductions ( pethanol ¼ 0.10 bar) andoxidations (xed at 500 Cand pwater ¼ 0.25 bar) in order to determine the redox decay ca-pacity of the iron oxide after each reduction and oxidationcycle.

F H 2

max.0.05 bar

0 10 20 30 40 50 60 70

0

1

2

3

4

5

6

F H 2

max.0.15 bar

F H 2

max.0.10 bar

0.050.100.15

M o

l a r

f l o w

( m

m o

l / m i n )

Time (min)

P Eth.(bar) F H 2 F H 2 O

Fig. 4 e Effect of the variation of ethanol partial pressure onthe reduction of fresh iron oxide at 675 C.

0 5 10 15 20 25 30 35 40 45 500.0

0.2

0.4

0.6

0.8

1.21.62.0

725ºC 700ºC

675ºC 650ºC

750ºC

625 ºC 650 ºC 675 ºC 700 ºC 725 ºC 750 ºC

H 2

O m o

l a r

f l o w ( m m o

l / m i n )

Time (min)

625ºC

Fig. 5 e Effect of temperature variation on the temporalevolution of water molar ow during the reduction step( pethanol [ 0.1 bar).

0 10 20 30 40 50 60 70 80 900.0

0.5

1.0

1.5

2.0

2.5

3.0

625 ºC 650 ºC 675 ºC 700 ºC 725 ºC 750 ºC

H 2

m o

l a r

f l o w

( m m o

l / m i n )

Time (min)

T

t*

T

F H2 max.

Fig. 6 e Inuence of reduction temperature on hydrogenproduced during the subsequent oxidation step withsteam at 500 C.

0.6

0.7

0.8

0.9

500 550 600 650 700 750 800

500 550 600 650 700 750 800

0.4

0.5

0.6

0.7

0.8

0.9

W G S

H 2

/ ( H 2

+ H

2 O ) ( a d i m )

Fe

FeO

Fe 3O4

(a)

StageCycle "B" "C" 1st

2 nd

3 rd

FeO

Fe

W G S

Fe 3O4

(b)

C O / ( C O + C O

2 ) ( a d i m )

Temperature(°C)

Fig. 7 e Baur e Glaessner and Water Gas Shift theoreticalequilibrium diagram and (a) ratio H 2 /(H2 D H2 O) and (b)ratio CO/(CO D CO2 ) in the exit gas stream along threesuccessive redox cycles.








Results are shown in Fig. 7 where ratios corresponding to aBaure Glaessner diagram [15] were calculated. This diagramexposes the existing thermodynamic equilibrium betweensolid phases depending on the partial pressures of the reducer(H2 and CO) and oxidant (CO 2 and H 2O) species. This way, thepertinent H 2 /(H2 þ H2O) (Fig. 7(a)) and CO/(CO þ CO2) (Fig. 7(b))ratios were obtained for both magnetite e iron (Stage “B” inFig. 1) and ethanol catalytic decomposition equilibria ( Stage“C” in Fig. 1). A similar behavior was found between H 2 /(H2 þ H2O) and CO/(CO þ CO2) ratios. Experimental ratios at

tested temperatures are indicated over the theoretical curves.Ratios corresponding to magnetite e iron equilibrium ( Stage“B”) are represented with black symbols, whereas the onesrelated to ethanol catalytic decomposition ( Stage “C”) aresymbolized by white symbols. There is no superposition of theratios obtained along Stage “B” over any of the curves exhibi-ted in the diagram. The reason for this behavior is thatreduction from Fe 3O4 to Fe through FeO consists ofa solid e gasreaction with slow kinetics compared to that of gas e gas re-actions, so that apparently the gas stream residence time inthe bed is not long enough to reach the equilibrium. Thus, thereducing potential of the gas stream leaving the reactor ishigher than that corresponding to the equilibrium condition

(i.e., higher H 2 and CO contents than actual). On the contrary,ethanol decomposition equilibrium perfectly overlaps withthe predicted equilibrium for the Water Gas Shift reaction(r.15), clearly implying that this reaction controls the compo-sition of the gases exiting the reactor once the iron oxide hastransformed in metallic iron during the reduction step.

On the other hand, time needed to oxidize the solid isfound to increase with the number of redox cycles. What ismore, a slight degradation of the structure due to the opera-tion at highreduction temperatures is added to the mentionedthermal ageing because of conducting experimental series of successive redox cycles: the higher the number of cycle, themore distant is the gas composition respecting that of the

magnetite e iron equilibrium ( Fig. 7).

3.7. Optimal condition parametric study

Various factors must be considered in order to decide theoptimal reduction temperature among those tested(625e 750 C). Hydrogen average production rate increases as itdoes the reduction temperature ( Fig. 5); therefore lowerethanol ow is necessary to completely reduce the solid athigher temperatures. In turn, the lower the reduction tem-perature, the higher the hydrogen production rate during thesubsequent oxidation step, probably due to a lower thermalstress. Taking into account the tradeoff between these oppo-site trends, a maximum is found around 675 C for hydrogenproduction rate per total cycle time (reduction e up to thebeginning of “ Stage C” e plus oxidation times) as shown inFig. 8. This graph exhibits a very similar average value of hydrogen production rate between 625 C and 675 C for therst redox cycle, nevertheless the production rate is clearlyhigher at 675 C for both, second and third cycles, in theoperation at 675 C. Thus, the optimum behavior regarding temperature is more clearly identied in subsequent cyclesafter the rst one. Values for 2nd cycle at 725 C and 2nd and3rd cycles at 750 C are not available (N.A. in Fig. 8) due to theintense pressure drop experimented by the solid bed by cokedeposition.

Although coke is formed during the reduction step, itadopts an inert role during oxidation steps at 500 C, keeping the purity of the hydrogen stream released. Consequently, theperiodic inclusion of a decoking step in the process should beneeded in order to remove periodically the deposited carbonin a separate stream apart from that where hydrogen isobtained.

4. Conclusions

The tested oxide (“triple” oxide based in hematite, aluminaand ceria) has proved to be able to decompose an ethanolstream at temperatures above 625 C. Main gaseous productsreleased from this decomposition (hydrogen and carbonmonoxide) can reduce the oxide in three marked stages: “A”,from hematite to magnetite consuming lattice oxygen fromthe iron oxide; “B”, from magnetite to metallic iron (probablythrough wustite although it was never detected), a stagemostly governed by a pseudo-equilibrium between solidphases and gas atmosphere, and nally a third stage “C”,clearly governed by the Water Gas Shift equilibrium. Thesestages have been checked out in several conditions (temper-atures and ethanol partial pressures). A hypothetical reactionnetwork that describes appropriately the reduction processhas been identied as well.

Coke formation has also been evidenced during thereduction step following an almost linear trend along time.Once the iron oxide has been completely reduced, reoxidationto magnetite by steam at 500 C releases hydrogen of highpurity with CO x species content lower than 1 ppm. This factallowsconcluding that at these temperatures, coke behaves asan inert material for steam gasication.

Also the effect of consecutive redox cycles has been tested,concluding that reduction temperature has a signicant effect

625 650 675 700 725 7500.0

0.1

0.2

0.3

0.4

0.5

N / A

N / A

A v e r a g e

H 2

p r o

d u c

t i o n r a

t e ( m m o

l / m i n )

Reduction temperature (ºC)

1st cycle 2nd cycle 3rd cycle

N / A

Fig. 8 e Average hydrogen production rate per cycle(reduction D oxidation time) for 3 consecutive redox cyclesand different temperatures.








in the behavior of the subsequent steps: the higher the tem-perature, the greater the sintering of the solid, but also thelower the content of carbon deposited on its surface. Thus, atradeoff between temperature and coke deposition along cy-cles has concluded that an optimum temperature of 675 Cshould be employed.

Acknowledgments

Financial support for this research has been provided by theSpanish Ministerio de Economı ´a y Competitividad (MINECO),through project ENE2010-16789. J. Plou also thanks the sameinstitution for the grant BES-2011-045092. Financial aid for themaintenance of the consolidated research group CREG hasbeen provided by the Fondo Social Europeo (FSE) through theGobierno de Aragon(Aragon, Spain).

r e f e r e n c e s

[1] Li W, Wang H, Ren Z, Wang G, Bai J. Co-production of hydrogen and multi-wall carbon nanotubes from ethanoldecomposition over Fe/Al 2O3 catalysts. Appl Catal B Environ2008;84(3e 4):433e 9.

[2] Jørgensen H, Kristensen JB, Felby C. Enzymatic conversion of lignocellulose into fermentable sugars: challenges andopportunities. Biofuels Bioprod Biorening 2007;1(2):119 e 34.

[3] Global Renewable Fuels Alliance (GRFA). Global ethanolproduction to reach 85.2 billion liters in 2012. World Grain;August 2012 .

[4] Balat M, Balat H. Recent trends in global production andutilization of bio-ethanol fuel. Appl Energy

2009;86(11):2273e

82.[5] Evans G, Smith C. 5.11-Biomass to liquids technology. In:Ali S, editor. Comprehensive renewable energy. Oxford:Elsevier; 2012. pp. 155e 204.

[6] Mattos LV, Jacobs G, Davis BH, Noronha FB. Production of hydrogen from ethanol: review of reaction mechanism andcatalyst deactivation. Chem Rev 2012;112(7):4094 e 123.

[7] Chaubey R, Sahu S, James OO, Maity S. A review ondevelopment of industrial processes and emerging

techniques for production of hydrogen from renewable andsustainable sources. Renew Sustain Energy Rev2013;23(0):443e 62.

[8] Mezalira DZ, Probst LD, Pronier S, Batonneau Y, Batiot-Dupeyrat C. Decomposition of ethanol over Ni/Al 2O3

catalysts to produce hydrogen and carbon nanostructuredmaterials. J Mol Catal a Chem 2011;340(1 e 2):15e 23.

[9] Hacker V, Fankhauser R, Faleschini G, Fuchs H, Friedrich K,Muhr M, et al. Hydrogen production by steam-iron process. JPower Sources 2000;86(1 e 2).

[10] Otsuka K, Kaburagi T, Yamada C, Takenaka S. Chemicalstorage of hydrogen by modied iron oxides. J Power Sources2003;122(2).

[11] Adanez J, Abad A, Garcia-Labiano F, Gayan P, deDiego LF. Progress in chemical-looping combustion andreforming technologies. Prog Energy Combust Sci2012;38(2):215e 82.

[12] Hu B, Yamaguchi Y, Fujimoto K. Low temperature methanolsynthesis in alcohol solvent over copper-based catalyst.Catal Commun 2009;10(12):1620 e 4.

[13] Pen ˜ a JA, Lorente E, Romero E, Herguido J. Kinetic study of theredox process for storing hydrogen reduction stage. CatalToday 2006;116(3):439e 44.

[14] Jozwiak WK, Kaczmarek E, Maniecki TP, Ignaczak W,Maniukiewicz W. Reduction behavior of iron oxides inhydrogen and carbon monoxide atmospheres. Appl Catal aGen 2007;326(1):17e 27.

[15] Baur E, Glaessner A. Equilibrium between iron oxides andcarbon monoxide and carbon dioxide. J Chem Soc1903;84(2):423.

[16] Lorente E, Pen ˜ a JA, Herguido J. Separation and storage of hydrogen by steam-iron process: effect of added metals uponhydrogen release and solid stability. J Power Sources2009;192(1):224e 9.

[17] Campo R, Dura n P, Plou J, Herguido J, Pen ˜ a JA. Combinedproduction and purication of hydrogen from methanolusing steam iron process in xed bed reactor. J PowerSources 2013;242:520 e 6.

[18] Kirchnerova J, Alifanti M, Delmon B. Evidence of phasecooperation in the LaCoO 3e CeO2e Co3O4 catalytic system inrelation to activity in methane combustion. Appl Catal a Gen2002;231(1e 2):65e 80.

[19] Lorente E, Pen ˜ a JA, Herguido J. Kinetic study of the redoxprocess for separating and storing hydrogen: oxidationstage and ageing of solid. Int J Hydrogen Energy2008;33(2):615e 26.


http://refhub.elsevier.com/S0360-3199(14)00051-2/sref1
















































































































































































































Documents

ethanol 5ad aafsdfastfwqfas sacaxvava axcvvvvvvvvvvvvvvvvasddddddddddd