hydrogen synthesis from biomass pyrolysis with in situ

Upload: albert-rc

Post on 04-Jun-2018

217 views

Category:

Documents


0 download

TRANSCRIPT

  • 8/13/2019 Hydrogen Synthesis From Biomass Pyrolysis With in Situ

    1/14

    Hydrogen synthesis from biomass pyrolysis with in situ

    carbon dioxide capture using calcium oxide

    Meilina Widyawati a, Tamara L. Church a, Nicholas H. Florin a,b, Andrew T. Harris a,*a Laboratory for Sustainable Technology, School of Chemical and Biomolecular Engineering, University of Sydney, NSW, 2006, Australiab Grantham Institute for Climate Change, Imperial College London, South Kensington Campus, London SW7 2AZ, United Kingdom

    a r t i c l e i n f o

    Article history:

    Received 23 August 2010

    Received in revised form

    14 November 2010

    Accepted 26 November 2010

    Available online 2 March 2011

    Keywords:

    Hydrogen

    Carbon dioxide capture

    Pyrolysis

    Pine

    CaOThermogravimetric-mass spectro-

    metric analysis

    a b s t r a c t

    Hydrogen (H2) and other gases (CO2, CO, CH4, H2O) produced during the pyrolysis of

    cellulose, xylan, lignin and pine (Pinus radiata), with and without added calcium oxide

    (CaO), were studied using thermogravimetry-mass spectrometry (TG-MS) and thermody-

    namic modeling. CaO improved the H2 yield from all feedstocks, and had the most

    significant effect on xylan. The weight loss of and gas evolution from the feedstocks were

    measured over the temperature range 150e950 C in order to investigate the principle

    mechanism(s) of H2formation. Without added CaO, little H2was produced during primary

    pyrolysis; rather, most H2 was generated from tar-cracking, reforming, and char-decom-

    position reactions at higher temperatures. When CaO was added, significant H2 was

    produced during primary pyrolysis, as the water-gas shift reaction was driven toward H2formation. CaO also increased the formation of H2 from reforming and char gasification

    reactions. Finally, CaO increased the extent of tar cracking and char decomposition, and

    lowered their onset temperatures. The production of H2 from pine over the course ofpyrolysis could be modeled by summing the H2 evolutions from the separate biomass

    components in relevant proportions.

    Copyright 2010, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights

    reserved.

    1. Introduction

    The concept of a hydrogen energy economy has attracted

    attention from the scientific, environmental, and political

    communities. The properties of H2 make it a promisingchemical feedstock, fuel, energy carrier and energy-storage

    medium[1]. Currently, about 96% of global H2is derived from

    fossil fuels [2], either through catalytic steam reforming,

    partial oxidation, or a combination of the two [1]. Producing H2from these finite fuel resources has raised concerns about

    energy security and about environmental sustainability, as

    the processes collectively liberate 0.4 Gt(CO2)/a [3]. Vastly

    preferable would be to produce H2from a renewable resource,

    such as water or biomass. H2can be obtained from water by

    electrolysis, thermolysis, or biologically assisted photolysis, or

    from biomass via pyrolysis or steam gasification (pyrolysis in

    the presence of added H2O)[4e7].

    Pyrolysisuses heat to degradea feedstock, in the absenceofan external oxidizing agent, to a combustible mixture of gases

    [1,8] that contains mainlyH2,CO,CO2, CH4 andH2O, along with

    tarand charfractions (Table 1,eq.(1)) [6]. The processoccursin

    three stages: (i) pyrolysis to produce the main gas products,

    volatiletar, andsolid char;(ii) crackingand reformingof tarand

    gases, and (iii) char gasification [7]. During each stage, the

    primary products participate in multiple, interdependent

    secondary reactions that can alter the nature and distribution

    * Corresponding author. Tel.:61 2 9351 2926; fax: 61 2 9351 2854.E-mail address:[email protected](A.T. Harris).

    A v a i l a b l e a t w w w . s c i e n c e d i r e c t . c o m

    j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c om / l o c a t e / h e

    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 6 ( 2 0 1 1 ) 4 8 0 0 e4 8 1 3

    0360-3199/$ e see front matter Copyright 2010, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.

    doi:10.1016/j.ijhydene.2010.11.103

    mailto:[email protected]://www.sciencedirect.com/http://www.elsevier.com/locate/hehttp://dx.doi.org/10.1016/j.ijhydene.2010.11.103http://dx.doi.org/10.1016/j.ijhydene.2010.11.103http://dx.doi.org/10.1016/j.ijhydene.2010.11.103http://dx.doi.org/10.1016/j.ijhydene.2010.11.103http://dx.doi.org/10.1016/j.ijhydene.2010.11.103http://dx.doi.org/10.1016/j.ijhydene.2010.11.103http://www.elsevier.com/locate/hehttp://www.sciencedirect.com/mailto:[email protected]
  • 8/13/2019 Hydrogen Synthesis From Biomass Pyrolysis With in Situ

    2/14

  • 8/13/2019 Hydrogen Synthesis From Biomass Pyrolysis With in Situ

    3/14

    sawdust[32]. Engineering aspects of steam gasification in the

    presence of CaO have also been considered [7,33]. The pyrolysis

    reactionsofhemicelluloseandlignininthepresenceofCaOhave

    not been studied, nor have the effects of various biomass

    components on one another during CaO-assisted pyrolysis. To

    address some of these gaps in understanding, we performed

    multi-technique analyses of the pyrolysis behaviours, with and

    without CaO present, of a biomass source (P. radiata, hereafterreferred to as pine) and three biomass constituents: cellulose,

    xylan (a hemicellulose), and lignin. This allowed us to study and

    compare the behaviours of each feedstock during pyrolysis, and

    to evaluate the effect of CO2 sorption by CaO on each, while

    focusing specifically on the generation of H2gas. We were also

    abletoconsidertheformationofH2 frompineinthecontextofH2production from its components.

    2. Materials and methods

    2.1. Feedstocks

    Avicel cellulose (Fluka), xylan from beechwood (Sigma) and

    lignin (organosolv, mixture of lignins from maple (50%), birch

    (35%), and poplar (15%), Aldrich) were selected as model

    biomass components. Pine wood (P. radiata, from Pollards

    Sawdust Supplies, Australia) was selected as a representative

    biomass sample, as pine products are commonly studied in

    biomass-gasification studies [15,17,21,31,32,34,35]. Pine does,

    however, pyrolyze differently from some hardwoods, such as

    holm oak and eucalyptus [34]. The as-received feedstocks

    were ground using a mortar and pestle, and sieved to 53/

    212 mm and CaCO3 (Aldrich) was sieved to

  • 8/13/2019 Hydrogen Synthesis From Biomass Pyrolysis With in Situ

    4/14

    the range 1e150. The lower heating rate was chosen for these

    experiments to facilitate the evolution of low-molecular-

    weight compounds [36]. The ions identified in this process

    were targeted for MS scanning during the pyrolysis experi-

    ments in order to obtain better resolution. Focusing on the

    detection of lighter species was a trade-off. Though it limited

    our ability to directly observe the effects of CaO on tar

    cracking, it allowed the evolution of light gases to be studied

    more accurately. By considering the equations in Table 1,

    insights on tar cracking can still be gleaned from the forma-

    tion of light gases, primarily H2and CH4.

    We examined several methods of reporting ion intensities

    [27,36,37]. Following three identical experiments on the

    pyrolysis of each biomass component, we found the best

    reproducibility using the relative integrated peak linear

    intensity to total integrated peak linear intensities normalized

    to sample weight, eq.(13).

    RInti Inti=X

    Inti m

    (13)

    where R(Int)i is the relative integrated peak linear intensity

    (hereafter referred to as relative intensity), Inti is the inte-grated intensity of a gas species, and m is themass of biomass.

    Though generally reproducible, this method of processing

    MS data returned highstandard deviations forsome speciesin

    some reactions, and this limited our ability to make definitive

    statements in some cases. However, TG-MS remains the only

    technique available to simultaneously measure the thermal

    decomposition and H2 output of a very small sample, and

    a small sample is necessary to minimize the interference of

    heat and mass transfer effects. Nevertheless, we cannot rule

    out heat and mass transfer as sources of error in these

    experiments.

    The prominent species for study were selected based on

    their relative intensities across the temperature range150e950 C and on their relevancy. Ion fragments with R

    (Int) < 0.5 nA min/mg were not considered prominent. Five ion

    species stood out as the most intense and relevant. The first of

    these, atm/e 2, is assigned to H2. The peak atm/e 15 shows

    the release of methyl groups, eCH3, which are formed

    primarily from methane. The parent H2O peak atm/e 18 was

    studied rather than the fragment peak atm/e 17. CO is the

    major contributor to the peak at m/e 28, though contribu-

    tions from other compounds, especially CO2and ethylene, are

    also likely [38]. Finally, the peak at m/e 44 indicates the

    release of CO2; acetaldehyde is a possible minor contributor to

    this peak. The most prominent minor ion species are given in

    the supporting data (Table S2).

    3. Results and discussion

    3.1. Predicted impact of CaO

    We have considered the importance of the equilibria inTables

    1 and 2 on biomass pyrolysis, and used thermodymanic

    modeling to study the effect of CaO on the steam gasification

    of cellulose [27]. For the present work, we used similar

    calculations to predict the impact of CaO on the products ofthe pyrolysis reactions of cellulose, xylan, lignin, and pine.

    These studies are for comparison purposes only; in a real

    system, biomass pyrolysis does not reach equilibrium for

    several reasons. At low temperatures, equilibrium is not

    approached because most of the material remains trapped in

    the biomass sample, and because the kinetic energy of the

    system is insufficient. The greater kinetic energy present at

    higher temperatures allows the reaction to come closer to

    equilibrium; however, it remains under kinetic control, as

    evidenced by the effect of catalysts on the reaction [39].

    Moreover, at hightemperatures, the reaction will divergefrom

    the particular equilibrium calculations described in this

    article because some components have been selectivelyremoved from the system (i.e. the ones that were volatilized

    and swept away at low temperature). Nevertheless, the ther-

    modynamic calculations offer insight into the effect of CaO on

    biomass pyrolysis.

    Our calculations, which are discussed in more detail in

    the supporting data, predict that CaO will increase the

    molar fraction (dry basis) of H2 in the pyrolysis products of

    all four biomass sources when T < 720 C (see Figs. S1eS4).

    At higher temperatures, the release of CO2from CaCO3 (the

    reverse of eq. (10) is favoured and CO2is no longer absorbed,

    thus negating the thermodynamic effects of CaO. The H2concentrations were maximized at 640 C. CaO (nCa/nC 0.5)

    increased the theoretical maximum H2concentration in theproduct gas produced from the pyrolysis of cellulose, xylan,

    lignin, and pine by 52, 49, 27, and 44%, respectively. Thus

    the effect of CaO on H2 concentration was inversely

    proportional to the C/O and H/O ratios of the feedstocks (see

    Fig. S5), and was therefore predicted to be greater in the

    pyrolysis of the polysaccharide biomass components than

    of lignin.

    Aside from altering the thermodynamic balance of gaseous

    products during pyrolysis, CaO catalyzes tar cracking[21e23].

    Further, even a small amount of carbonate salt affects the

    initial products of rapid cellulose pyrolysis quite significantly

    [17]. The CaO added to our samples, and the Ca(OH) 2 and

    CaCO3 that are formed in situ, may catalyze one or more

    Table 4 e Composition of ash formed from the combustion of xylan and pine.a

    Ash source Content in ash sample [%]

    SiO2 Al2O3 Fe2O3 TiO2 K2O MgO Na2O CaO SO3 CO2 P2O5

    Xylan 0.5 0.5 0.2 0.03 0.8 0.6 51.8 6.5 0.5 34.5 7.30

    Pine 4.0 1.3 24.0 0.09 12.6 15.7 2.0 37.8 2.1 0.2b 0.17

    a Ash was analyzed using ICP-OES after combusting to 815C, fusing with borate, and dissolving in acid.b Calculated by difference.

    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 6 ( 2 0 1 1 ) 4 8 0 0 e4 8 1 3 4803

    http://dx.doi.org/10.1016/j.ijhydene.2010.11.103http://dx.doi.org/10.1016/j.ijhydene.2010.11.103http://dx.doi.org/10.1016/j.ijhydene.2010.11.103http://dx.doi.org/10.1016/j.ijhydene.2010.11.103
  • 8/13/2019 Hydrogen Synthesis From Biomass Pyrolysis With in Situ

    5/14

    reactions during pyrolysis; however, we are unable to distin-

    guish kinetic and thermodynamic effects in our experiments.

    Thus, in the discussions below,we will consider the formation

    of various products within the context of the reactions shown

    inTables 1 and 2. The catalytic effects of calcium compounds

    will not be discussed, but they are possible whenever these

    are present.

    3.2. Pyrolysis of biomass and components

    The main thermogravimetric features and the total

    ion intensities for major ions in the pyrolyses of cellulose,

    xylan, lignin and pine are summarized in Table 5, and the

    temperatures at which various ion productions are maxi-

    mized are listed inTable S3.

    3.2.1. Cellulose

    Even beforethe gasification of biomasswas commonly studied

    asasourceofH2, cellulosepyrolysis wasstudied with thegoals

    of understanding fires [40] and deriving chemicals from

    biomass [41]. Cellulose samples are more uniform than

    hemicellulose, lignin, or biomass samples [42,43], making

    them easier to study. Thus the pyrolysis of cellulose is betterunderstood than those of other biomass components, and the

    kinetics of the process have been reviewed [44]. The ther-

    mogravimetric behaviour of cellulose is also well-known

    [14e16,45e48]. It generally begins with a large, sharp weight

    loss withTmaxw355 C; the preciseTmaxdepends on the heat-

    ing rate [11e14,40]. This is followed by a smaller, more gradual

    weight loss at Tw650e700 C. Little char remains after pyrol-

    ysis. We observed a large, sharp weight loss with Tmax 362 C

    and a much smaller feature that peaked at 705 C(Fig. 1a). H2production was slow and relatively constant for T

  • 8/13/2019 Hydrogen Synthesis From Biomass Pyrolysis With in Situ

    6/14

    a feature at 415 C in the thermal decomposition of cellulosewith added Ca(OH)2 [15]; however, much less calcium was

    used in that case (nCa/nC 0.02e0.05). In any case, H2production declined during this shoulder, reaching

    a minimum near 435 C, then slowly increased to a broad

    peak at 580 C. The rate of mass loss was quite low in the

    first half of this broad peak, suggesting that H 2 was mainly

    produced by hydrocarbon reformation and tar cracking (eqs.

    3e5). This is supported by an increase in intensity of the ion

    at m/e 15 (Tpeak 500 C), which could correspond to CH4formation during tar cracking (eq. (5)). Some of the CH4produced during tar cracking could then be consumed to

    produce H2 (eq. (3)). These reactions occurred at higher

    temperatures when no CaO was added (Fig. 1a).

    Above 500 C, as the detection of methyl groups decreased

    but H2production continued to increase, the rates of weight

    loss and CO formation also began to increase (Tpeak 680 C

    for weight loss). Thus char decomposition (eqs. 6e8) likely

    began around this temperature. No char was detected at the

    end of cellulose pyrolysis in the presence of CaO. CO2production also increased during this period, likely from both

    char (eq. (7)) and carbonate (reverse of eq. (10)) decomposition.Both processes contributed to the weight loss. The decompo-

    sitions of carbonate and char are not independent; the

    increased pressure of CO2formed in the former may favour

    char decomposition via the Boudouard reaction (eq. (8)),

    accounting for some of the CO formed (though its contribution

    is expected to be limited below 800 C); whereas the increase

    in CO2 pressure from carbonate decomposition disfavours

    char decomposition. The wateregas reactions (eqs. 6 and 7)

    may also have contributed to weight loss, as they had in the

    absence of CaO (vide supra); this is also consistent with the

    coincident peak in CO formation. CO may also have been

    formed by tar cracking, which is catalyzed by CaO[21e23].

    Overall, CaO appeared to increase the relative intensity ofH2 formation over the course of cellulose pyrolysis (Table 5,

    entries 1 and 2), consistent with the results of our experiments

    under steam gasification conditions [14]. In this case,

    however, the high uncertainty in H2 production from cellulose

    alone (entry 1) precludes a statistically significant comparison.

    3.2.2. Xylan

    There are fewer pyrolysis studies of hemicellulose than

    cellulose[50], though thermogravimetric analyses have been

    reported for the pyrolysis reactions of hemicelluloses under

    inert atmosphere[45,46,48,51]and under syn gas and H2[11].

    Consistent with previous studies, the thermal decomposition

    we observed for xylan (Fig. 2a) occurred chiefly between 200and 330 C and left 24 wt% char at 850 C. In addition to the

    largest weight loss peak (Tmax 304 C), we observed

    a shoulder atT 250 C. This is likely due to the specific xylan

    used, and may indicate that the sample was inhomogeneous.

    Though several studies have examined the thermal decom-

    position of xylan, only one has measured the H2 released

    during pyrolysis under inert atmosphere. Yang et al. used

    GC-TCD to follow gas production from xylan pyrolysis in

    a fixed bed system, but were not able to measure the weight

    loss simultaneously [48]. Bymeasuring H2 formation and mass

    loss concurrently, we found that xylan produced little H2during its primary thermal decomposition (200e330 C);

    rather, H2O was the main gas detected during this period. Asecond (but unresolved) weight-loss peak occurred with

    a maximum at 430 C. In addition to continued H2O release,

    this weight loss was associated with minor peaks in the

    production of CO (Tpeak 470 C) and CO2(Tpeak 430 C) that

    were likely due to the continued breakdown of the biomass

    sample. The wateregas oxidations of carbon (eqs. 6 and 7) are

    also possible sources of carbon oxides, but are likely minor

    contributors in this temperature range.

    Following the production of CO and H2O during the weight

    loss with maximum at 430C, H2and methyl-group formation

    increased, signaling the onset of the second stage of the

    reaction. At first, the H2and methyl peaks grew concurrently,

    suggesting that they originated from the same source; this

    Fig. 1 eThe rate of weight loss and the release of major

    gases during the pyrolysis of (a) cellulose and (b) cellulose

    in the presence of CaO, as measured by TG-MS.

    vT/vt[ 40 C/min. MS data is presented in arbitrary units,

    and the curve for H2was multiplied by 20 to improve

    visibility. In (b), nCa/nC [ 0.5.

    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 6 ( 2 0 1 1 ) 4 8 0 0 e4 8 1 3 4805

    http://dx.doi.org/10.1016/j.ijhydene.2010.11.103http://dx.doi.org/10.1016/j.ijhydene.2010.11.103http://dx.doi.org/10.1016/j.ijhydene.2010.11.103http://dx.doi.org/10.1016/j.ijhydene.2010.11.103
  • 8/13/2019 Hydrogen Synthesis From Biomass Pyrolysis With in Situ

    7/14

    was most likely tar cracking (eq. (5)). The evolution of CH4peaked at 495 C and then declined, whereas H2 production

    continued to increase (Tpeak 535 C) even as the rate of mass

    loss continued to decrease. Thus, atT > 495 C, CH4was likely

    consumed by steam reformation to produce H2 and CO (eq.

    (3)). Accordingly, CO production began to increase around

    550 C, and continued to rise as the char began to decompose,

    producing a final peak in the rate of mass loss at 755 C.

    AddingCaO hada significant effecton thepyrolysis ofxylan

    (Fig. 2b). Thermal decomposition began much as the reaction

    without CaO had, with mass loss beginning at the same

    temperature (Tinit 195 C), and H2O being evolved. Again,

    some H2O was likelytaken upby CaO (eq.(11)).The weight loss

    peakedat294 C,nearpeaksinCOandCO2 evolution.However,

    in contrast to the reaction without added CaO, H 2production

    also peaked during this time. H2 formation increased as CO

    production decreased, suggesting that the wateregas shift

    reaction (eq. (2)) was important during this stage of the reac-

    tion. Immediately following the first large weight loss, anotherwas observed. This weight loss had a maximum at 435 C, and

    was primarily associated with H2O evolution. The H2O

    production in this case likely derived from Ca(OH)2via dehy-

    dration (the reverse ofeq. (11)) orcarbonation(eq.(12)). Therate

    of weight loss in this case (0.15 %wt/C) was consistent with

    that seen for cellulose (0.14 %wt/C; Fig. 1b) at the same

    temperature. Towardthe end of the second large mass loss, H2production declined, then increased sharply (Tpeak 525 C).

    The increasedH 2 production wasaccompanied by increases in

    methyl-group and CO production, consistent with tar cracking

    (eq. (5)). Some CH4and CO may also have been reacted to form

    H2via the steam-gasification (eq. (3)) and wateregas shift (eq.

    (2)) reactions, respectively. H2production decreased slowly asthetemperatureincreasedfurther,and a thirdweightlosspeak

    occurred, reaching a maximum at 690 C. This peak was asso-

    ciated with CO and CO2 release, and is assigned to the

    decompositions of CaCO3 (the reverse of eq.(10)) andchar (eqs.

    (6)e(8)); tar-cracking reactions likely also contribute to the CO

    and CO2release. Evincing char decomposition, only 3% char

    remained after xylan pyrolysis in the presence of CaO (cf. 24%

    whenCaOwasnotadded).Overall,CaOincreasedH2 formation

    from xylan by 165%.

    3.2.3. Lignin

    The pyrolysis behaviours of a wide range of lignins have been

    studied [11,12,17,45,46,48,52e56]. Contrary to the morechemically uniform polysaccharide components of biomass,

    lignins are generally pyrolyzed over a broad temperature

    range under inert atmosphere [12,45,46,48,52e56]and under

    H2and syngas[11]. Both the different chemical compositions

    of lignins from different sources [42,43,50,52,55,56] and the

    different methods of lignin extraction[17,50,54,56]contribute

    to the variation among the published data. The thermogravi-

    metric data we measured for lignin (Fig. 3a) was generally

    consistent with that in previous reports; decomposition

    occurred over wide temperature range (215e825 C;

    Tmax 408 C), and produced 29 wt% char at 850 C.

    Compared to cellulose and xylan, lignin produced a higher

    relative intensity of methyl groups, but a lower relativeintensity of H2O (see Table 5, entries 1, 3, and 5). It also

    produced a higher relative intensity of minor fragments (see

    Table S2). Several studies have examined the products of

    lignin pyrolysis [12,17,46,48,52,53,55], and a few have moni-

    tored H2 evolution in the reaction. Unlike Yang et al. (who

    studied commercial alkali lignin) [48] and Fushimi et al.

    (commercial lignin) [12], we observed a small peak in H2production at Tmax (408 C). Faix et al. observed maxima at

    440 C (i.e. slightly aboveTmax) in the pyrolysis of lignins from

    beech and bamboo woods [55]. In addition to a peak in H2formation, we also observed peaks in the formation of CO,

    methyl groups, H2O, and CO2at this temperature, indicating

    that primary pyrolysis was occurring. Several studies have

    Fig. 2 eThe rate of weight loss and the release of major

    gases during the pyrolysis of (a) xylan and (b) xylan in the

    presence of CaO, as measured by TG-MS. vT/vt[ 40 C/

    min. MS data is presented in arbitrary units, and the curve

    for H2was multiplied by 10 to improve visibility. In (b),

    nCa/nC [ 0.5.

    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 6 ( 2 0 1 1 ) 4 8 0 0 e4 8 1 34806

    http://dx.doi.org/10.1016/j.ijhydene.2010.11.103http://dx.doi.org/10.1016/j.ijhydene.2010.11.103http://dx.doi.org/10.1016/j.ijhydene.2010.11.103http://dx.doi.org/10.1016/j.ijhydene.2010.11.103
  • 8/13/2019 Hydrogen Synthesis From Biomass Pyrolysis With in Situ

    8/14

    observed maxima in these gases atTmax[12,48,52], though the

    relative sizes of these peaks varied among the reports. Thus

    most lignins appear to undergo primary pyrolysis at similar

    temperatures, but with varying product distributions. MS

    studies by Faix et al. found peaks in the release of several

    oxygenated species, including HCHO, CHO, H3COH, and

    H3CO inthe 420e500 C range, but little ofthese species above

    540 C, and therefore concluded that most of the oxygen was

    removed during this phase[55].

    The greatest H2 production from lignin occurred at high

    temperatures. Previous studies observed the greatest rate of

    H2 release between 640 and 800 C, depending on the lignin

    source [12,48,55]. We found a broad peak in H2 production

    between 500 and 900 C (Tpeak 750 C), and peaks in CO and

    CO2formation at 700 C. The latter result was consistent with

    the observations of Yang et al. and Fushimi et al. [12,48],

    though others have reported declining CO and CO2 levels atthis temperature [52,55]. In the present system, the coinci-

    dent occurrence of the peaks in CO, CO2, and H2suggest that

    char decomposition via the wateregas reactions (eqs. 6 and 7)

    is the dominant process at 700 C, though the dehydrogena-

    tion of aromatic rings is also important [55]. Finally, we

    observed a decrease in H2formation at T > 750 C. This was

    followed by an increase in H2O detection, suggesting that

    some H2 may have been consumed to form water via the

    reverse of the wateregas shift reaction (i.e. the reverse of

    eq. (2)). However, the decreasing rate of CO detection at these

    temperatures eliminates the wateregas shift reaction as

    a major contributor to H2O production. Rather, the increase

    in H2O formation was likely caused by the primary pyrolysisof recalcitrant compounds, consistent with the coincident

    small peak in weight loss that ends as H2O generation

    decreases (T > 800 C).

    CaO did not affect the primary weight loss of lignin, which

    still peaked just above 400 C (Tmax 407 C;Fig. 3b). A new

    weight loss peak at 700 C, associated with carbonate and char

    decompositions, was also observed, just as it was in the

    pyrolysis reactions of cellulose and xylan in the presence of

    CaO. The release of H2O from Ca(OH)2may have occurred at

    w425 C, as it had in the cases of cellulose and xylan pyrolysis

    in the presence of CaO; however, no new peak was visible due

    to the peak in lignin decomposition in this temperature range.

    The pyrolysis of lignin in the presence of CaO produced 70%more H2 than that without CaO (Table 5, compare entries 5

    and 6), and yielded only 5% char (cf. 29% without CaO).

    In the absence of CaO, H2had been a minor product of the

    primary pyrolysis (i.e. near Tmax). With CaO, however, it had

    a significant peak in this temperature range. Thus CaO likely

    promoted the wateregas shift reaction (eq. (2)) during the

    primary pyrolysis of lignin, as it had during the primary

    pyrolysis reactions of cellulose and xylan. H2 production

    slowed to a minimum at Tw500 C, then began to increase

    (Tpeak635 C,withashoulderat710 C)evenastheweightloss

    rate of the sample stayed was relatively stable. Considered

    together with the concurrent decrease in the relative intensity

    ofCH4, this suggeststhatthe steam reforming of methane andhydrocarbons (eqs 3 and 4), promoted by the removal of CO2(and consequently CO) by CaO, were likely important at this

    temperature. The rate of H2 formation continued to increase as

    the weight loss rate increased, suggesting that these processes

    continued while the more stable lignin compounds, as well as

    char and carbonate, decomposed.

    3.2.4. Pine

    The gasifications of various pine species have been examined

    [15,17,21,31,32,34,35] and the thermal decompositions of two

    pine specieshave been considered in terms of cellulose, xylan,

    and lignin pyrolysis [35]. In the present study, the thermal

    decomposition of pine (Fig. 4a) occurred in the range

    Fig. 3 eThe rate of weight loss and the release of major

    gases during the pyrolysis of (a) lignin and (b) lignin in the

    presence of CaO, as measured by TG-MS. vT/vt[ 40 C/

    min. MS data is presented in arbitrary units, and the curve

    for H2was multiplied by 10 to improve visibility. In (b),nCa/nC [ 0.5.

    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 6 ( 2 0 1 1 ) 4 8 0 0 e4 8 1 3 4807

    http://dx.doi.org/10.1016/j.ijhydene.2010.11.103http://dx.doi.org/10.1016/j.ijhydene.2010.11.103http://dx.doi.org/10.1016/j.ijhydene.2010.11.103http://dx.doi.org/10.1016/j.ijhydene.2010.11.103
  • 8/13/2019 Hydrogen Synthesis From Biomass Pyrolysis With in Situ

    9/14

    200e530 C, with the greatest weight loss being completed

    whileT < 425 C. The major pyrolysis peak (Tmax 388 C) was

    unsymmetrical, with a large shoulder at lower temperatures,

    as previously discussed in detail by Grnli et al.[35]. H2O, CO2,

    and CO were the major gaseous products of primary pyrolysis,

    though methyl groups were also detected in this temperature

    range (Tpeak 385 and 440 C, overlapping). Contrary to the

    pyrolysis reactions of its components, the thermal decompo-

    sition of pine did not show a high-temperature peak. Never-

    theless,the pyrolysis of pine produced lesschar (at 850 C, 15%

    of the original weight of pine remained) than the thermal

    decompositions of xylan and lignin, but more than that of

    cellulose.

    Except for a very small, sharp peak at 385 C, the rate of H2formation from pine increased steadily over the course of the

    pyrolysis, reaching a maximum at 750 C. Similarly to the

    reactions with biomass components, the pyrolysis of pine

    formed the greatest relative intensity of H2when the rate of

    weight loss was low. Thus most of the H 2 formed from pinecame from secondary reactions. These were likely the

    wateregas shift reaction (eqs. (2)) at lower temperatures, and

    hydrocarbon reforming (eqs 3 and 4) and wateregas reactions

    (eqs 6 and 7) at higher temperatures.

    Having examined the formation of H2from pine, we turned

    our attention toward the reaction in the presence of added

    CaO. Though the pyrolysis of pine [15,34] and the steam

    gasifications of pine sawdust[32,34]and bark[31]have been

    studied in the presence of CaO, H2 production from pine

    pyrolysis in the presence of CaO has not. We found that added

    CaO had little effect on Tinit and Tmax of thermal pine degra-

    dation (seeFig. 4b andTable 5, entries 7 and 8), but added the

    same two features that had appeared when the componentswere pyrolyzed in the presence of added CaO: a low shoulder

    (T 412 C) associated with the evolution of H2O (at approxi-

    mately the same weight loss rate that was observed for

    cellulose and xylan) and an additional weight loss (maximum

    at 680 C) associated with CO and CO2 release (char and

    carbonate decomposition, tar cracking).

    H2 became an important product of primary pyrolysis

    when CaO was present. Toward the end of the rapid weight

    loss, H2formation declined as the rate of weight loss declined,

    but it increased again after the temperature passed 450 C. The

    detection of methyl groups peaked at 510 C and declined as

    that of H2increased, implying that steam reformation (eqs. 3

    and 4) was occurring. H2formation continued to increase asCO and CO2production increased again, suggesting that char

    decomposition by the wateregas reactions was producing H2.

    The maximum relative intensity of H2came at 660 C. Overall,

    the relative intensity of H2increased by 82% in the presence of

    added CaO.

    The pine and component studies indicate that adding CaO

    can significantly increase the H2yield from biomass pyrolysis,

    decrease char yields, and lower the temperature at which the

    relative intensity of H2 is maximized. A comparison of the

    thermogravimetric and gas-evolution data allows some

    general deductions to be made about which reactions are

    most affected by CaO. At lower temperatures (T < 450 C), the

    increased relative intensity of H2and complete suppression ofCO2reveal the importance of the exothermic wateregas shift

    reaction (eq. (2)). This was especially visible in carbohydrate

    (cellulose or xylan) pyrolysis. For all feedstocks, added CaO

    also lowered the Tpeak for methyl-group detection. This indi-

    cates enhanced tar cracking (eq. (5)) at lower temperatures,

    though we cannot determine whether this is due to kinetic

    factors (catalysis by CaO) or thermodynamic changes(the heat

    released from carbonation and the shifts in the equilibria that

    govern the system). Regardless, increased tar cracking

    releases more methane and small hydrocarbons that can

    undergo steam reformation (eqs 3 and 4). At moderate

    temperatures (500 < T < 700 C), pyrolysis in the presence of

    CaO produces more H2 at the expense of methyl-group

    Fig. 4 eThe rate of weight loss and the release of major

    gases during the pyrolysis of (a) pine and (b) pine in the

    presence of CaO, as measured by TG-MS. vT/vt[ 40 C/

    min. MS data is presented in arbitrary units, and the curve

    for H2was multiplied by 20 to improve visibility. In (b), nCa/

    nC [ 0.5.

    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 6 ( 2 0 1 1 ) 4 8 0 0 e4 8 1 34808

    http://dx.doi.org/10.1016/j.ijhydene.2010.11.103http://dx.doi.org/10.1016/j.ijhydene.2010.11.103http://dx.doi.org/10.1016/j.ijhydene.2010.11.103http://dx.doi.org/10.1016/j.ijhydene.2010.11.103
  • 8/13/2019 Hydrogen Synthesis From Biomass Pyrolysis With in Situ

    10/14

    formation, supporting the involvement of steam-reforming

    reactions (eqs 3 and 4). This is most noticeable in the pyrolysis

    of lignin, which gives high relative intensity of methyl groups.

    CaO also increased the extent, and decreased the onset

    temperature, of char decomposition. This resulted in higher

    H2 production at higher temperatures (T > 600 C), and

    decreased the amount of char left after the reaction. In future

    work, it would be interesting to consider the influence ofheating rate and residence time on these effects, as the

    reactions involved in pyrolysis may be quite different under

    very rapid heating or in a bed reactor.

    Though added CaO increased H2 production from all

    feedstocks (the difference was not statistically significant in

    the case of cellulose), the effect is by no means optimized. We

    were not able to judge the conversion of CaO to CaCO3, so it is

    possible that not all of the added CaO was available to absorb

    CO2. We were also unable to evaluate the degree to which char

    deposition on the reactive surface of CaO hindered CO2sorp-

    tion; this could seriously impact CaO activity in a large reactor

    system, and requires study before these results can be applied

    to large-scale biomass-gasification systems. Further, forpractical reasons (see Materials and Methods), we used

    nCa/nC 0.5 in these experiments, but more CaO may benefit

    the pyrolysis of biomass even more; for example, we have

    previously shown that nCa/nC 1 is ideal for H2 production

    from cellulose[27].

    3.3. Component additivity

    There is ongoing discussion in the literature regarding

    whether biomass pyrolysis can be understood as the sum of

    the pyrolysis behaviours of its components, i.e., whether

    a simple additivity relationship applies between the pyrol-

    ysis of cellulose, hemicellulose, and lignin [57]. To evaluatethis idea, several groups have pyrolyzed mixtures of biomass

    components, and compared the results to those of the pyrol-

    ysis reactions of individual components or of natural biomass

    samples. Caballero et al. reported that the thermal decom-

    position of a mixture of lignin and holocellulose (a mixture of

    cellulose and hemicellulose in relevant proportions) did not

    replicate that of almond shells [58]. Yang et al. studied the

    thermal decomposition of cellulose, xylan, and lignin, as well

    as of mixtures of two or three components, and reported

    negligible interaction among them [45]. For samples with

    known cellulose, hemicellulose, and lignin components, they

    were able to predict the fraction of sample weight that would

    be lost in three different temperature ranges, though theconverse prediction was less reliable. Couhert and co-workers

    determined that even the method of mixing impacted the CO2yields obtained in the flash pyrolysis of mixtures of cellulose,

    hemicellulose, and lignin [42,43]. Wang et al. [11]pyrolyzed

    sawdust, cellulose, hemicellulose and lignin under syn gas

    and under H2, and found that the presence of lignin and

    hemicellulose affected the thermogravimetric behaviour of

    cellulose. Evans and Milne concluded that the organic

    components of biomass did not impact one another during

    flash pyrolysis, but that the inorganic species present in real

    biomass samples impacted their pyrolysis [17]. Hosoya and

    co-workers compared the flash pyrolysis of untreated and

    demineralized Japanese Oak samples, and concluded that the

    mineral species in wood could not fully account for the

    differences between product distributions from biomass and

    component mixtures [59]. Nowakowski and Jones, on the

    other hand, found that HCl treatment had little effect on the

    thermogravimetric behaviour of cellulose and lignin [46].

    However, chemical treatments are known to impact the

    pyrolysis behaviours of biomass components [17], making

    definitive conclusions from these studies difficult.Aside from the interactions between biomass components

    and the presence of inorganic salts, several other factors

    complicate the modeling of biomass pyrolysis using compo-

    nent additivity. For example, lignins vary among species

    [42,43,50,52,55,60], as do hemicelluloses [42,43,50,60]. In fact, it

    is not trivial to find a component that is representative of

    hemicellulose [61]. Moreover, real biomass samples contain

    low-molecular-weight organic extractives [18] that are also

    pyrolyzed. Finally, the problem is compounded by the varia-

    tion in methods for determining the composition of a biomass

    sample[62]. Nevertheless, many authors have attempted to

    use mathematical models to relate the pyrolysis behaviour of

    biomass to the separate pyrolysis behaviours of biomasscomponents. In general, models that describe thermogravi-

    metric behaviour[18,35,61e64] are more common and, thus

    far, more successful than models that aim to predict product

    compositions [42,43,58,62]. However, models that accurately

    predict thermogravimetric behaviour can require the use of

    6e10 parameters. Better results are obtained when these

    parameters relate to the weight loss of a sample in different

    temperature ranges rather than its actual cellulose, hemi-

    cellulose, and lignin contents[18,45,63]. Among models that

    attempt to predict the formation of gases from biomass,

    Caballero et al. found good agreement between the predicted

    and experimental yields of CO, CO2, and H2O from almond

    shells, although those of CH4, methanol, formaldehyde, andhigher hydrocarbons were not accurately predicted. Couhert

    et al. showed that physically meaningful results were not

    obtained when building models based on the H2, CO, and CH4yields from the rapid pyrolysis of hardwood, softwood, wood

    bark, rice husk, and grass [42]. To our knowledge, no study has

    predicted the H2formation profile of a biomass sample, and

    we therefore examined this possibility using the pyrolysis of

    pine, with and without added CaO.

    As thermogravimetric models of biomass pyrolysis are

    more common than gas-formation models, we first attempted

    a quantitative comparison of the thermal decomposition of

    pine with those of its components by combining the cellulose,

    xylan, and lignin traces in the ratio 4:3:3 [65]. The results,shown in the supporting data (Figs. S6 and S7for pine and

    pine/CaO, respectively), were unimpressive. The predicted

    and experimental traces showed some similarities, but the

    simple weighted sum of the component results was not an

    accurate predictor of the thermal decomposition of pine. This

    is likely due to the simplicity of the model, as Grnli et al.

    produced much better models for the thermogravimetric

    behaviour (for T 150e450 C) of two pine species upon

    optimizing ten parameters[35].

    In the component studies, H2 formation was not neces-

    sarily proportional to the rate of thermal decomposition;

    rather, it was often at a maximum when the rate of thermal

    decomposition was low. Thus, despite the failure of

    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 6 ( 2 0 1 1 ) 4 8 0 0 e4 8 1 3 4809

    http://dx.doi.org/10.1016/j.ijhydene.2010.11.103http://dx.doi.org/10.1016/j.ijhydene.2010.11.103http://dx.doi.org/10.1016/j.ijhydene.2010.11.103http://dx.doi.org/10.1016/j.ijhydene.2010.11.103
  • 8/13/2019 Hydrogen Synthesis From Biomass Pyrolysis With in Situ

    11/14

    component additivity to predict the thermal decomposition of

    pine, we examined the relationships between H2 formation

    during pine pyrolysis and during the pyrolysis of the biomass

    components. The results are shown in Figs. 5 and 6 for

    pyrolysis in the absence and presence of CaO, respectively.

    The H2data were collected as relative intensity, and the data

    from the three different components were normalized (to the

    same maximum R(Int)) before comparison, so quantitative

    conclusions are inappropriate. Qualitatively, though, the

    weighted sum of H2 formation from each of the three

    component traces was a surprisingly accurate predictor of the

    experimentally observed trace (Figs. 5a and 6a). Most notably,

    the separate component traces predicted the temperature at

    Fig. 5 e H2production during the pyrolysis of pine. MS data

    is presented with arbitrary units. vT/vt[ 40 C/min. (a)

    Predicted and experimental H2production. The predicted

    profile was calculated as the weighted sum of the

    normalized component decompositions according to: R

    (Int)pine [ 0.4 R(Int)cellulose D0.3 R(Int)xylan D0.3 R(Int)lignin.

    (b) H2production during the pyrolysis of pine compared to

    H2production from the biomass components.

    Fig. 6 eH2production during the pyrolysis of pine in the

    presence of CaO. MS data is presented with arbitrary units.

    nCa/nC [ 0.5; vT/vt[ 40 C/min. (a) Predicted and

    experimental H2production. The predicted profile was

    calculated as the weighted sum of the normalizedcomponent decompositions according to: R(Int)pine [ 0.4 R

    (Int)cellulose D0.3 R(Int)xylan D0.3 R(Int)lignin. (b) H2production during the pyrolysis of pine in the presence of

    CaO, compared to H2production from the biomass

    components in the presence of CaO.

    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 6 ( 2 0 1 1 ) 4 8 0 0 e4 8 1 34810

    http://dx.doi.org/10.1016/j.ijhydene.2010.11.103http://dx.doi.org/10.1016/j.ijhydene.2010.11.103http://dx.doi.org/10.1016/j.ijhydene.2010.11.103http://dx.doi.org/10.1016/j.ijhydene.2010.11.103
  • 8/13/2019 Hydrogen Synthesis From Biomass Pyrolysis With in Situ

    12/14

    which the relative intensity of H2was maximized; this is an

    important parameter if biomass pyrolysis is to be used for H2synthesis.

    2The agreement between the predicted and experimental

    H2-evolution curves provides some license to consider H2formation from the components independently. A compar-

    ison of the H2 formation from pine with those from the

    biomass components (Fig. 5b), suggests that, in the earlystages of the reaction, H2was likely produced by the pyrolysis

    of cellulose, as the cellulose and pine traces were very similar

    forT < 300 C. Lignin began to contribute to H2formation just

    beforeT 400 C. Though xylan undergoes thermal decom-

    position at a lower temperature than cellulose and lignin do,

    it releases little H2at low temperatures. It was thus unlikely

    to be an important contributor to H2 formation until just

    below 500 C. The contribution from lignin pyrolysis

    increased significantly shortly thereafter, and continued to

    increase even as the overall H2 production declined. This is

    consistent with the results of Demirbas, who showed that the

    H2 yields from pyrolysis were higher for cellulose and

    hemicellulose at lower temperatures, but higher for ligninabove 675 C[66]. The relative intensity of H2was maximized

    at 730 (experimental) and 720 C (predicted). When CaO was

    added, cellulose remained the main source of H2 during

    primary pyrolysis, although xylan and lignin also contributed

    in this case (Fig. 6b). All three components contributed to H2formation at higher temperatures, and the maximum relative

    intensity of H2 came at 660 (experimental) and 620 C

    (predicted).

    H2 formation during pine pyrolysis can be qualitatively

    modeled using the H2formation traces from the pyrolyses of

    cellulose, xylan, and lignin, both in the absence and pres-

    ence of CaO. This suggests that the H2production from pine

    can be understood as the simple weighted sum of H2productions from these components, despite that the ther-

    mogravimetric behaviour of pine is modeled less accurately

    using this strategy. This result, taken together with the

    observation that H2formation is often high at temperatures

    when the rate of weight loss is low, reaffirms the impor-

    tance of gas-phase reactions in determining the product

    distribution obtained from biomass pyrolysis. It also

    suggests that component additivity may be useful for pre-

    dicting the formation of H2 from other sources of biomass.

    The results obtained by Couhert et al. [42,43], however,

    illustrate an important caveat to these predictions: compo-

    nent additivity may not be useful when the biomass and

    components are from very different sources. These authorsobtained very similar H2 yields in the flash pyrolysis of

    beechwood and of a mixture of spruce and fir, but found

    that mathematical models based on these two samples plus

    the pyrolysis data from grass, wood bark, and rice husk gave

    unreasonable results. Thus component additivity models

    such as the ones shown in Figs. 5a and 6a, which were

    developed using hemicellulose and lignin from wood sour-

    ces, may only be useful for predicting the behaviour of

    woods. More investigations will be necessary before we

    can properly evaluate the scope of this model; however,

    we hope that the results of this will be useful for the

    further development of advanced biomass-gasification

    technologies.

    4. Conclusion

    The combination of thermogravimetric and mass-spectro-

    metric analyses offered insight into H2 formation from

    biomass pyrolysis. Although significant relative errors were

    obtained forthe total relative intensities of the species studied

    in repeated experiments, the changes in gas formation withtemperature were quite reproducible. Thus comparing the

    evolution profiles of the various gases, along with the rate of

    mass loss, allowed conclusions to be drawn about the domi-

    nant processes operating during the reaction. Most of the H2generated from biomass pyrolysis is actually produced from

    secondary reactions among CO, CO2, H2O and CH4, and

    between these gases and heavier products.

    CaO is known to improve the H2 yield from biomass

    pyrolysis by acting as a CO2 sorbent, andwe have examined its

    specific actions during the reaction. Adding CaO increased H 2production from several of the gas-phase and gasesolid

    reactions that influence H2formation. At lower temperatures,

    the H2 yield from the wateregas shift reaction greatlyincreased when CO2was being absorbed. As the temperature

    was raised, CaO encouraged the cracking of tars, and drove

    the steam reformations of methane and hydrocarbons to

    produce more H2. Finally, CaO lowered the onset temperature

    of char decomposition, and decreased the char yield. The

    result was more H2produced from the wateregas reactions.

    We also found that the H2 evolution from pine could be

    qualitatively predicted from the weighted sum of the

    component H2-evolution traces, despite that the thermogra-

    vimetric behaviour was not predicted as accurately. This was

    true both with and without added CaO. These results allow the

    synthesis of H2from pine to be considered in the context of its

    components, and may be applicable to other biomass sources.

    Acknowledgements

    The authors thank EON and the Australian Research Council

    for their support of this research through an EON Research

    Initiative Grant and DP666488, respectively. M.W. is grateful to

    the Asian Development Bank and the University of Sydney

    World Scholars Programme for scholarships.

    Appendix

    Supplementary data related to this article can be found online

    atdoi:10.1016/j.ijhydene.2010.11.103.

    r e f e r e n c e s

    [1] Rand DAJ, Dell RM, editors. Hydrogen energy: challenges andprospects. Royal Society of Chemistry; 2008.

    [2] IEA. Oil supply security: the emergency response of IEAcountries in 2007. International Energy Agency; 2007.

    [3] Pacala S, Socolow R. Stabilization wedges: solving theclimate problem for the next 50 years with current

    technologies. Science 2004;305:968e72.

    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 6 ( 2 0 1 1 ) 4 8 0 0 e4 8 1 3 4811

    http://dx.doi.org/10.1016/j.ijhydene.2010.11.103http://dx.doi.org/10.1016/j.ijhydene.2010.11.103http://dx.doi.org/10.1016/j.ijhydene.2010.11.103http://dx.doi.org/10.1016/j.ijhydene.2010.11.103http://dx.doi.org/10.1016/j.ijhydene.2010.11.103http://dx.doi.org/10.1016/j.ijhydene.2010.11.103
  • 8/13/2019 Hydrogen Synthesis From Biomass Pyrolysis With in Situ

    13/14

    [4] (a) Holladay JD, Hu J, King DL, Wang Y. An overview ofhydrogen production technologies. Catal Today 2009;139:244e60;(b) Levin DB, Chahine R. Challenges for renewable hydrogenproduction from biomass. Int J Hydrogen Energy 2010;35:4962e9.

    [5] (a) Kalinci Y, Hepbasli A, Dincer I. Biomass-based hydrogenproduction: a review and analysis. Int J Hydrogen Energy

    2009;34:8799e817;(b) Nath K, Das D. Hydrogen from biomass. Curr Sci(Bangalore) 2003;85:265e71;(c) Navarro RM, Sanchez-Sanchez MC, Alvarez-Galvan MC,del Valle F, Fierro JLG. Hydrogen production from renewablesources: biomass and photocatalytic opportunities. EnergyEnviron Sci 2009;2:35e54.

    [6] Ni M, Leung DYC, Leung MKH, Sumathy K. An overview ofhydrogen production from biomass. Fuel Process Technol2006;87:461e72.

    [7] Florin NH, Harris AT. Enhanced hydrogen production frombiomass with in situ carbon dioxide capture using calciumoxide sorbents. Chem Eng Sci 2008;63:287e316.

    [8] McKendry P. Energy production from biomass (part 2):conversion technologies. Bioresour Technol 2002;83:47e54.

    [9] Shafizadeh F. Introduction to pyrolysis of biomass. J AnalAppl Pyrolysis 1982;3:283e305.

    [10] McKendry P. Energy production from biomass (part 1):overview of biomass. Bioresour Technol 2002;83:37e46.

    [11] Wang G, Li W, Li B, Chen H. TG study on pyrolysis ofbiomass and its three components under syngas. Fuel2008;87:552e8.

    [12] Fushimi C, Araki K, Yamaguchi Y, Tsutsumi A. Effect ofheating rate on steam gasification of biomass. 2.Thermogravimetric-Mass Spectrometric (TG-MS) analysis ofgas evolution. Ind Eng Chem Res 2003;42:3929e36.

    [13] (a) Milosavljevic I, Oja V, Suuberg EM. Thermal effects incellulose pyrolysis: relationship to char formation processes.Ind Eng Chem Res 1996;35:653e62;(b) Milosavljevic I, Suuberg EM. Cellulose thermal

    decomposition kinetics: global mass loss kinetics. Ind EngChem Res 1995;34:1081e91.

    [14] Florin NH, Harris AT. Mechanistic study of enhanced H2synthesis in biomass gasifiers with in-situ CO2capture usingCaO. AIChE J 2008;54:1096e109.

    [15] Liodakis SE, Statheropoulos MK, Tzamtzis NE, Pappa AA,Parissakis GK. The effect of salt and oxide-hydroxideadditives on the pyrolysis of cellulose and Pinus halepensispine needles. Thermochim Acta 1996;278:99e108.

    [16] Varhegyi G, Antal Jr MJ, Szekely T, Till F, Jakab E.Simultaneous thermogravimetric-mass spectrometricstudies of the thermal decomposition of biopolymers. 1.Avicel cellulose in the presence and absence of catalysts.Energy Fuels 1988;2:267e72.

    [17] Evans RJ, Milne TA. Molecular characterization of the

    pyrolysis of biomass. Energy Fuels 1987;1:123e37.[18] Orfao JJM, Antunes FJA, Figueiredo JL. Pyrolysis kinetics of

    lignocellulosic materialsethree independent reactionsmodel. Fuel 1999;78:349e58.

    [19] Harrison DP. Sorption-Enhanced hydrogen production:a review. Ind Eng Chem Res 2008;47:6486e501.

    [20] Beruto DT, Botter R. Liquid-like H2O adsorption layers tocatalyze the Ca(OH)2/CO2solid-gas reaction and to forma non-protective solid product layer at 20C. J Eur Ceram Soc2000;20:497e503.

    [21] (a) Delgado J, Aznar MP, Corella J. Calcined dolomite,magnesite, and calcite for cleaning hot gas from a fluidizedbed biomass gasifier with steam: life and usefulness. Ind EngChem Res 1996;35:3637e43;(b) Delgado J, Aznar MP, Corella J. Biomass gasification with

    steam in fluidized bed: effectiveness of CaO, MgO, andCaO-MgO for hot raw gas cleaning. Ind Eng Chem Res 1997;36:1535e43.

    [22] Simell PA, Leppalahti JK, Bredenberg JBS. Catalyticpurification of tarry fuel gas with carbonate rocks andferrous materials. Fuel 1992;71:211e8.

    [23] Hanaoka T, Yoshida T, Fujimoto S, Kamei K, Harada M,Suzuki Y, et al. Hydrogen production from woody biomass by

    steam gasification using a CO2sorbent. Biomass Bioenergy2005;28:63e8.

    [24] http://webbook.nist.gov/chemistry/NIST Standard ReferenceDatabase Number 69.

    [25] Nan Z-D, Shi Z-Y, Qin M, Hou W-G, Tan Z-C. Formationprocess and thermodynamic properties of Calcite. Chin JChem 2007;25:592e5.

    [26] Florin N, Harris AT. Hydrogen production from biomass.Environmentalist 2007;27:207e15.

    [27] Florin NH, Harris AT. Hydrogen production from biomasscoupled with carbon dioxide capture: the implications ofthermodynamic equilibrium. Int J Hydrogen Energy 2007;32:4119e34.

    [28] Acharya B, Dutta A, Basu P. An investigation into steamgasification of biomass for hydrogen enriched gas production

    in presence of CaO. Int J Hydrogen Energy 2010;35:1582e9.[29] Hanaoka T, Fujimoto S, Yoshida T, Kamei K, Harada M,

    Suzuki Y, et al. Hydrogen production from woody biomass bynovel gasification using CO2sorbent. Stud Surf Sci Catal 2004;153:103e8.

    [30] Fujimoto S, Yoshida T, Hanaoka T, Matsumura Y, Lin S-Y,Minowa T, et al. A kinetic study of in situ CO2removalgasification of woody biomass for hydrogen production.Biomass Bioenergy 2007;31:556e62.

    [31] Mahishi MR, Goswami DY. An experimental study ofhydrogen production by gasification of biomass in thepresence of a CO2sorbent. Int J Hydrogen Energy 2007;32:2803e8.

    [32] (a) Hu G, Huang H. Hydrogen rich fuel gas production bygasification of wet biomass using a CO2sorbent. Biomass

    Bioenergy 2009;33:899e906;(b) Wei L, Xu S, Liu J, Liu C, Liu S. Hydrogen production insteam gasification of biomass with CaO as a CO2absorbent.Energy Fuels 2008;22:1997e2004.

    [33] (a) Acharya B, Dutta A, Basu P. Chemical-looping gasificationof biomass for hydrogen-enriched gas production with in-process carbon dioxide capture. Energy Fuels 2009;23:5077e83;(b) Mahishi MR, Sadrameli MS, Vijayaraghavan S,Goswami DY. A novel approach to enhance the hydrogenyield of biomass gasification using CO2sorbent. J Eng GasTurbines Power 2008;130. 011501/011501e011501/011508;(c) Pfeifer C, Puchner B, Hofbauer H. In-situ CO2-absorptionin a dual fluidized bed biomass steam gasifier to producea hydrogen rich syngas. Int J Chem React Eng 2007;5:A9.

    [34] Franco C, Pinto F, Gulyurtlu I, Cabrita I. The study ofreactions influencing the biomass steam gasificationprocess. Fuel 2003;82:835e42.

    [35] Grnli MG, Varhegyi G, Di Blasi C. Thermogravimetricanalysis and devolatilization kinetics of wood. Ind Eng ChemRes 2002;41:4201e8.

    [36] Jakab E, Till F, Varhegyi G. Thermogravimetric-massspectrometric study on the low temperature oxidation ofcoals. Fuel Process Technol 1991;28:221e38.

    [37] Arenillas A, Rubiera F, Pis JJ. Simultaneousthermogravimetric-mass spectrometric study on thepyrolysis behavior of different rank coals. J Anal ApplPyrolysis 1999;50:31e46.

    [38] The experiments were performed in Ar rather than N2, so thelatter does not contribute significant intensity at m/e 28.

    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 6 ( 2 0 1 1 ) 4 8 0 0 e4 8 1 34812

    http://dx.doi.org/10.1016/j.ijhydene.2010.11.103http://dx.doi.org/10.1016/j.ijhydene.2010.11.103http://dx.doi.org/10.1016/j.ijhydene.2010.11.103http://dx.doi.org/10.1016/j.ijhydene.2010.11.103
  • 8/13/2019 Hydrogen Synthesis From Biomass Pyrolysis With in Situ

    14/14