1
Au oca aly ic kine ic model o he mog a ime ic analysis and
composi ion es ima ion o biomass and polyme ic ac ions
A. Cabeza, F. Sob ón, F.M. Yed o, and J. Ga cía-Se na*
High P essu e P ocesses G oup, Depa men o Chemical Enginee ing and
En i onmen al Tech., Uni e si y o Valladolid, 47011 Valladolid, Spain
*Co esponding au ho : Tel.: +34 983184934
E-mail: jgse [email protected] a.es (J. Ga cía-Se na)
Abs ac
A comp ehensi e kine ic model o slow py olysis o biomass du ing a
The mog a ime ic analysis (TGA) has been de eloped, including he simula ion o
a iable hea ing a es, composi ion es ima ion and s uc u al analysis o biomass.
Biomass was assumed as a ma ix o h ee solid global componen s (hemicellulose,
cellulose and lignin) in which wa e and oil can be also p esen .
Kine ics we e based on an au o-ca aly ic model because i can simula e he
deg ada ion in cellulosic ma e ials, as he clea age o he biopolyme s p oduce
oligome s ha accele a e he u he depolyme isa ion. The eac ion pa hway ollowed
he Wa e loo’s mechanism, which s ablishes ha all solid compounds decompose in o
ola iles and cha coal. This mechanism was comple ed by he apo iza ion o wa e
and oil, and assuming ha he o med cha coal can b eak in o ola iles by a slow
eac ion. The se was sol ed by he 8 h Runge-Ku a’s me hod and alida ed by he
Simplex Nelde -Mead and B oyden-Fle che -Gold a b-Shanno’s me hods. The
de elopmen o his model has a high in e es because i can help o unde s and how
he con e sion om biomass o biochemicals akes place.
To assess wha pa ame e s can a ec he he mal deg ada ion o biomass pu e
polyme ic samples o hemicellulose, cellulose and lignin and complex samples (seeds
and woody biomass) we e s udied and i ed. Two ypes o ope a ions we e conside ed
oo. An iso he mal deg ada ion om 150ºC up o 350ºC wi h inc emen s o 50ºC, and a
non-iso he mal decomposi ion wi h hea ing a e o 5ºC/min, 10ºC/min and 20ºC/min up
o empe a u es a ound 800ºC, depending on he s udied sample. A e age absolu e
de ia ions lowe han 7% we e ob ained. I was deduced ha he e a e some
in e ac ions be ween he h ee main biomass compounds. These in e ac ions we e
obse ed by he a ia ions in he kine ic pa ame e s be ween complex and pu e
2
samples, also hey we e pe cei ed be ween he iso he mal and no-iso he mal way. On
he o he hand, an e ec o he biomass s uc u e has been epo ed by he di e ences
be ween he kine ics o he seeds and o he woody samples. I is ema kable ha he
de eloped model could ep oduce he cellulose decomposi ion wi h a a iable hea ing
a e using a unique se o kine ic pa ame e s. This was possible by a no-A henius’
dependence wi h empe a u e. In he same way, i was used o p edic he ini ial
composi ion o he s udied biomass wi h de ia ions lowe han 7% o lignin and
cellulose.
Keywo ds: Au oca aly ic kine ic, composi ion es ima ion, TGA, cellulose, hemicellulose,
lignin.
1. In oduc ion
The use o ossil uels as he main aw ma e ial o indus y is no sus ainable, and
ce ainly i will no be he o e e -solu ion. So a new sou ce o basic compounds (i.e.
ca bon, hyd ogen and oxygen) and ene gy should be conside ed. This new sou ce
could be biomass [1], which can be ans o med in o bioene gy, biochemical and
bio uels in bio e ine ies [2, 3]. Howe e , he design o hese bio e ine ies equi es
knowledge abou he con e sion om aw ma e ial o uels and as , cheap and
accu a e biomass-analysing me hods. Fo he la e , se e al we me hods o chemical
analysis ha e been used [4]. These me hods a e based on he ac iona ion o
biomass samples and a la e isola ion o pu i ied ac ions, which could be quan i ied
using con en ional analy ical ins umen s. Al hough hese echniques ha e high
accu acy and obus ness, hey a e no sui able o an indus ial scale because hey a e
expensi e and equi e a lo o ime. Ano he op ion would be spec oscopic analysis,
such as, he Nea In a ed Re lec ance (NIR) spec oscopy, which educes ime
equi emen s and cos and i is a me hod wi h a high ep oducibili y. Ne e heless,
hese analysis need da a wi h a e y high quali y and an ini ial blank spec um, which is
an impo an limi a ion. So, he measu emen o he ini ial biomass composi ion is an
issue ha ha e no an op imal solu ion ye . The mog a ime ic analysis (TGA) o
biomass could be he answe o his p oblem unde ce ain condi ions. In addi ion, i
can p o ide in o ma ion abou how he he mal decomposi ion akes place.
The mog a ime ic analysis is a slow py olysis p ocess which consis o eco ding he
mass a ia ion o a sample which is ea ed wi h a empe a u e p o ile. This p o ile is
p o ided by a gas phase which can be an ine o an oxidan compound [5]. This ype
o analysis ha e been s udied ho oughly [5-7] and he e a e a lo o wo ks abou hei
3
modelling in he li e a u e. The mos ex ended model conside s a i s o de kine ic o
each compound p esen in biomass assuming ha biomass is o med by h ee main
compounds (cellulose, hemicellulose and lignin). These componen s decompose o
cha coal and ola iles by independen eac ions. S. Völke [8] used a i s o de kine ic
o adjus he decomposi ion o pu e cellulose and he de ia ion be ween he
expe imen al da a and he simula ion was ela i ely high. In con as , Capa R e al .
[9] s udied he pu e cellulose he mal b eaking bu conside ing an au oca aly ic model
which supposed a good i ing wi h an o e all de ia ion a ound 1 %. On he o he hand,
V. Mangu e al. [10] p oposed a kine ic model o n h-o de o he deg ada ion o
esidues om oma o p ocessing indus y which could ep oduce he biomass
beha iou . Bu A. Zabanio oua e al. [11], K. Slopiecka e al. [12] and E. Kas anaki e
al. [13] s udied he TGA kine ics o se e al lignocellulosic biomass samples, popla
wood and ligni e-biomass blends espec i ely wi h a i s o de model and hey
ob ained good i s oo. K. Slopiecka e al. and A. Zabanio oua e al. i ed hei TGA as
a single compound, which is use ul o ep oduce he decomposi ion. Howe e , i is no
capable o ep oducing he indi idual beha iou o he biomass componen s and
ensu ing ha he ob ained pa ame e s ha e physical meaning. On he o he hand, E.
Kas anaki e al. and V. Mangu e al. adjus ed hei TGA wi h indi idual kine ics o
each biomass compound. The e o e, he beha iou o each o hem could be simula ed
and he physical sense o he pa ame e s checked. Ne e heless, hey did no s udied
he causes o he a ia ions in he kine ics o he biomass componen s assuming ha
he e a e no in e ac ions be ween hem. Taking in o accoun his big ange o possible
models i is di icul o selec one because any o hem could be a good way o simula e
he he mal deg ada ion o biomass. Finally, he au oca aly ic model is he op ion
selec ed in his wo k due o he ac ha i can ep oduce he s eep changes in
cellulosic ma e ial be e han a i s o n h-o de model.
The aim o his pape is o s udy he main pa ame e s ha a ec o he he mal
decomposi ion o lignocellulosic biomass in a he mog a ime ic analysis de eloping a
model which could ep oduce he decomposi ion o any sample o biomass and i s
componen s in an iso he mal o non-iso he mal p ocess and a any hea ing a e. The
las one is impo an because a disc epancy in he amoun o o med cha coal has
been epo ed by o he au ho s [8, 9, 12, 14, 15] when di e en hea ing a es a e used.
Fu he mo e, he causes o he changes in he kine ics o he indi idual biomass
componen s wi h he ype o samples o p ocess was assessed by he compa ison o
he kine ics pa ame e s, which is no gene ally done in p e ious s udies. In addi ion,
his model should be able o es ima e he ini ial composi ion o he deg aded sample
4
om he he mog a ime ic analysis and om he kine ic pa ame e s i ed p e iously.
This capaci y is impo an because i is a new use o he TGA modelling and, i i is
de eloped co ec ly, i would become an economic op ion o ob aining he ini ial
composi ion o he biomass.
2. Ma e ials and me hods
2.1. Ma e ials
G ape seeds om Vi is ini e a L (Temp anillo) om Ma a ome a S.A. wine y
(Valbuena de Due o, Spain) campaign 2011 and se e al woody was es we e used as
aw ma e ial. Ma e ial o his s udy was g ound o a pa icle size o 0.5-1.0 mm.
The eagen s used o HPLC analysis we e: cellobiose (+98%), glucose (+99%),
uc ose (+99%), glyce aldehyde (95%), py u aldehyde (40%), a abinose (+99%), 5-
hyd oxyme hyl u u al (99%), lac ic acid (85%), o mic acid (98%), ac ylic acid (99%),
mannose (+99%), xylose (+99%) and galac ose (+99%) pu chased om Sigma and
used wi hou u he modi ica ion. Fo he s uc u al ca bohyd a es and lignin
de e mina ion sul u ic acid (98%) and calcium ca bona e (≥ 99.0%) we e used as
eagen s supplied by Sigma oo.
2.2. Biomass cha ac e iza ion
2.2.1. Suga con en
The suga con en measu emen o a biomass sample equi es i s hyd o he mal
ac iona ion ollowed by a hyd olysis o he p oduc which will be ed o a HPLC la e .
The hyd olysis is needed because he ac iona ion gene a es a ange o polyme ic
ac ions and compounds which could no be di ec ly iden i ied in a HPLC. So hey
ha e o be clea aged by a u he hyd olysis in o hei basic uni s o monome s, e.g.
glucose, uc ose, xylose and a abinose.
The samples we e hyd olysed adding 3.00±0.01 ml o sulphu ic acid (72%) o 15 ml o
each aliquo . Each sample was incuba ed in a o ced con ec ion o en o 30±5 min a
30±3ºC. A e his ime, he samples we e aken ou om he o en, hey we e dilu ed by
84.00±0.04 ml o deionized wa e and inally hey we e placed in he o en o 1 hou a
121 ºC. A e wa ds, he solu ion was cooled down o oom empe a u e and i was
il e ed unde acuum. Be o e injec ing in he HPLC he samples we e neu alized o
pH=6-7 using calcium ca bona e.
5
The HPLC column used o he sepa a ion o he compounds was a SUGAR SH-1011
Shodex a 50 ºC a a low o 0.8 ml/min using a solu ion o 0.01N o sulphu ic acid and
Milli-Q wa e as mobile phase. A Wa e s IR de ec o 2414 and Wa e s dual λ
abso bance de ec o 2487 (210 nm and 254 nm) we e used o iden i y he suga s and
hei de i a i es.
2.2.2. Solid analysis. Klason lignin de e mina ion and suga s a ached o he solid
The aw ma e ial and he solid esidue gene a ed by he hyd olysis we e analysed o
lignin con en using he Klason assay acco ding o he TAPPI s anda d me hod T-222
om-98 [16]. To do so, 300 mg o sample was pu in o labo a o y glass bo les, 3 mL o
sul u ic acid (72%) was added and i was incuba ed du ing 30 min a 30ºC and i was
shaken igo ously e e y 5-10 min. Then, he mix u e was dilu ed wi h 84 mL o
deionized wa e and i was placed in an o en o 1 h a 121ºC. A ha momen , he
sample was aken ou om he o en, cooled down o oom empe a u e and he
mix u e was il e ed unde acuum. The ob ained solid a e il a ion was d ied a
105ºC o 24 h, i was cooled down in a desicca o and hen i was weigh ed. This solid
was in oduced in he calcina ion o en a 550ºC o 24 h o de e mine he ash con en .
Conside ing he weigh di e ences, he Klason lignin con en was calcula ed. The
hyd olysis liquid was neu alized wi h calcium ca bona e o pH=6-7, hen i was il e ed
and analysed by HPLC as explained in sec ion 2.3.1.2 Suga s.
The ini ial composi ion calcula ed by he me hods desc ibed in 2.2.1 and 2.2.2 is
collec ed in ¡E o ! No se encuen a el o igen de la e e encia..
2.3. Expe imen al se -up and p ocedu e
TGA we e ca ied ou in a TGA/SDTA RSI analyze o Me le Toledo. Samples o
app oxima ely 10 mg we e hea ed om 50ºC o he equi ed empe a u e a a a e o
20ºC/min unde N2 a mosphe e (60 NmL/min low) o de e mine he ca boniza ion. The
inal empe a u e changed wi h he ype o analysis. I he s udy was a iso he mal
condi ions i had a alue be ween 150ºC and 350 ºC. Howe e , when i was no
iso he mal he biomass we e hea ed up o empe a u es a ound 800 ºC.
2.3.1. P ocedu e o he analysis o he e ec o he composi ion
The mog a ime ic analysis a a iable empe a u e wi h a hea ing a e o 20ºC/min o
woody samples wi h di e en lignin con en we e i ed. So he di e ence be ween he
kine ics pa ame e s we e used o disco e how he composi ion a ec s o he he mal
deg ada ion. TGA o hemicellulose, cellulose and g ape skins, which does no ha e
lignin, a he same hea ing a e we e pe o med o s udy his ac o oo.
2.3.2. P ocedu e o he analysis o he e ec o he s uc u e

6
The e ec o he biomass s uc u e was s udied compa ing he adjus ed kine ic
pa ame e s ob ained om he TGA (wi h a hea ing a e o 20ºC/min) o wo ypes o
pu e lignin: an alkaline lignin and a sample om Tu ku, Finland. The las one was
ex ac ed using a hyd o opic subs ance, he p- oluene sul ona e. In addi ion, he
de ia ion o he kine ics pa ame e s be ween he TGA (hea ing a e o 20ºC/min) o a
sample o g ape seeds and g ape seeds ex ac ed wi h a mix u e o e hanol/wa e
(70/30) o 1 hou was conside ed. The kine ics a ia ion be ween hese g ape seeds
and hyd olysed g ape seeds o 1 hou a h ee di e en empe a u es (250ºC, 300ºC
and 340 ºC) and a a hea ing a e o 20 ºC/min we e analysed oo.
2.3.3. P ocedu e o he analysis o he e ec o he hea ing a e
The ole o he hea ing a e was conside ed by i ing TGA o pu e cellulose a h ee
di e en hea ing a e: 5ºC, 10ºC, 20ºC and compa ing he alues o hei kine ics
pa ame e s.
2.3.4. P ocedu e o he analysis o he e ec o he iso he mal condi ions
This ac o was s udied by he adjus men o he TGA o Ace Saccha um, a ype o
maple, a 5 empe a u es (150ºC, 200ºC, 250ºC, 300ºC and 350ºC) wi h a hea ing a e
o 20 ºC/min and conside ing he modi ica ions in he kine ics.
3. Ma hema ical model
3.1. Biomass composi ion
The solid o ganic compounds p esen in biomass a e di ided in h ee main biopolyme
ac ions: hemicellulose, cellulose and lignin. Hemicellulose and cellulose a e
cons i u ed by monome ic suga s. The di e ence be ween hem is ha cellulose is a
linea polyme o anhyd oglucopy anose (hexose) uni s linked by e he bonds, while
hemicellulose is a b anched and amo phous polyme o med by bo h pen oses and
hexoses. In con as , lignin is a complex, c oss-linked, h ee-dimensional a oma ic
polyme o med wi h phenylp opane uni s [17]. The lowe he amoun o lignin he mo e
lexible is he biomass (e.g. he bs). In addi ion, some ine s (ino ganic mine al
compounds) and wo liquid phases can be p esen inside he biomass: wa e and an
o ganic phase which is iden i ied in his wo k as oil.
7
Simpli ying he aw biomass s uc u e we can conside ha he cellulose mic o ibe s
a e connec ed by hemicellulose in 3D s uc u e o lignin ha encloses and p o ec s
hem (Figu e 1). The s uc u e o he biopolyme ac ions and o he kinds biomass
lignin-lean can be a bi di e en .
Figu e 1 HERE
Table 1: HERE
3.2. Reac ion pa hway
Deg ada ion p ocess
The he mal deg ada ion o biomass in an ine a mosphe e (slow py olysis condi ions)
s a s wi h he apo iza ion o liquid phases. Wa e e apo a es nea 100ºC and oil
be ween 100ºC and 300ºC. In his esea ch, we in en ionally did no d y he biomass
un il ull d yness o mimic as much as possible some kind o humid condi ions and
he e o e he wa e e apo a ion.
A 200ºC lignin begins i s decomposi ion, b eaking i s weake pa s and enhancing he
eac ion o hemicellulose and cellulose. Be ween 250ºC and 275ºC hemicellulose
eac s and a ound 300 ºC i disappea s comple ely, which p omo es he cellulose
b eaking. The las one commences i s deg ada ion be ween 300ºC and 350ºC and,
om his poin , only lignin, ine subs ances and he p oduc om he decomposi ion
(cha coal) emains in he biomass. Lignin deple es a 500ºC and cha coal con inues in
he sample wi h a e y low deg ada ion a e. Cha coal only ades comple ely i he
a mosphe e is changed o an oxidan compound.
Reac ion mechanism
The mos ex ended idea in he li e a u e [5, 6, 9-13] is a eac ion pa hway based on he
Wa e loo-mechanism. This heo y es ablishes ha , unde slow py olysis condi ions,
biomass decomposes o cha coal and gases ia dehyd a ion eac ions [14, 15].
Following his heo y, i was assumed in his wo k ha any o ganic solid compound in
he biomass could be con e ed in o cha coal and ola iles by independen pa allel
eac ions. Fu he mo e, an indi idual cha coal o each solid o ganic componen was
used in his pape (Figu e 2), wi h he in en ion o desc ibing he slow py olysis p ocess
in de ail. This leads o a subs an ial imp o emen in he i ing o he expe imen al da a
8
be e u he p edic ion. The pa hway was comple ed adding he decomposi ion o
each cha coal o ola iles and wi h he apo iza ion o liquid phases i p esen .
Figu e 2: HERE
3.3. The model
Assump ions
Aimed a simpli ying he modelling p oblem i was assumed ha :
a. All he eac ions a e i e e sible and independen . So, he deg ada ion kine ics
o each componen only depend on hei composi ion and empe a u e [10, 12,
13].
b. The e a e no ene gy anspo limi a ions wi hin he biomass pa icles (as only
10 mg o mic onized pa icles we e used o he TGA analysis). Consequen ly,
all he pa s o he biomass a e a he same empe a u e.
c. Di usional mass anspo esis ances o liquid phases a e negligible, as he
pa icles we e mic onized.
Mass balances
The model o he decomposi ion p ocess conside ed a non-s a iona y mass balance o
each componen in he biomass sample:
𝑑𝑚𝑗
𝑑𝑡 =𝑟𝑗=∑𝑔𝑖𝑗·𝑟𝑖
𝑁𝑟
𝑖=1
( 1 )
And he o al a ia ion o mass was calcula ed by he addi ion o all o hem:
𝑑𝑀
𝑑𝑡=∑𝑑𝑚𝑗
𝑑𝑡
𝑁
𝑗=1
( 2 )
Kine ics
Mass a ia ion was caused by wo di e en phenomena: eac ion kine ics o he solid
ma e ial and mass ans e o he liquid phases.
 Liquid mass ans e : ee oil (o ex ac able compounds) and ee wa e .
9
Mass ans e o liquid phases was desc ibed by he pa ial mass ans e coe icien in
he gas phase, he mass ans e a ea and he di e ence be ween de equilib ium
concen a ion in he liquid phase and he global concen a ion in he gas phase (as
d i ing o ce) ( 3 ).
𝑟𝑖=ℎ·𝑆·(𝐶𝑗∗−𝐶𝑗)
( 3 )
As he ope a ing p essu e is he a mosphe ic, he equilib ium concen a ion ( 4 ) was
ob ained by he ideal gas equa ion and he apou p essu e calcula ed by he An oine
equa ion ( 5 ) o each compound. The An oine’s equa ion coe icien s o each liquid
phase a e compiled in Table 2.
𝐶𝑗∗=𝑃𝑗∗
𝑅·𝑇
( 4 )
ln(𝑃𝑗∗)=Aj+𝐵𝑗
𝐶𝑗+𝑇+𝐷𝑗·ln(𝑇)+𝐸𝑗·𝑇𝐹𝑗
( 5 )
Table 2: HERE
In addi ion, he ans e a ea was conside ed as a unc ion o he mass in he solid, so
he inal exp ession o he mass ans e was:
𝑟𝑖=ℎ·(𝐶𝑗∗)·𝑚𝑗𝑛𝑙𝑖
( 6 )
 Solid kine ics: o solid o ganic compounds
As i was men ioned in he in oduc o y sec ion, he e a e wo op ions o empe a u e
dependen kine ics, i.e. a i s o de eac ion ( 7 ) and an au oca aly ic eac ion ( 8 ).
The i s is he mos ex ended op ion in he bibliog aphy [9-13] and he second is
p oposed because i s esponse is e y simila o he beha iou o he biomass
obse ed in he li e a u e [9] and in p e ious s udies. Bo h kine ic equa ions conside ed
an A henius’ dependence wi h empe a u e.
𝑟𝑖=𝑘𝑖·𝑚𝑖=𝑘𝑜𝑖·𝑒−𝐸𝑎𝑖
𝑅·𝑇·𝑚𝑗
( 7 )
16
kine ics dec ease espec he non- ea ed samples. Howe e , hemicellulose kine ics
p esen s he opposi e beha iou bu i is no ep esen a i e because he main pa o
hemicellulose disappea du ing he hyd olysis.
Figu e 8: HERE
The lignin kine ic a ia ions obse ed in Figu e 8 could be jus i ied again wi h i s
p o ec ion unc ion like in pa 4.2.1.1. On he o he hand, cellulose shows in his case a
highe modi ica ion in kine ics due o hei deg ada ion du ing he hyd olysis. Finally,
empe a u e educe again he di e ences be ween he he mal deg ada ion o bo h
samples. K8 and K9 do no appea because o hei low con ibu ion o he o al cha .
4.2.2. Woody biomass
In his poin , he he mal decomposi ion in iso he mal and no iso he mal condi ions o
woody biomass is conside ed. All he i ings needed a modi ica ion in he kine ics
pa ame e s.
4.2.2.1. Non-iso he mal p ocess
Using he pa ame e s ob ained om he g ape seeds he wood deg ada ion could no
be di ec ly simula ed because i has a di e en s uc u e. The expe imen al da a we e
di ided in se e al g oups depending on i s lignin con en (Table 4) o can use he same
pa ame e s o each g oup (¡E o ! No se encuen a el o igen de la e e encia.).
This di ision shows ha a di e ence in he composi ion implies a change in he
decomposi ion oo. This could be explained by an in e ac ion be ween he species in
he wood, so when a g ea e amoun o lignin exi s he b eaking o cellulose in o
ola iles and i s cha coal o ma ion a e educed and he deg ada ion in o ola iles o
he la e oo. In addi ion, a highe p oduc ion o ola iles om lignin and deg ada ion
in o cha coal is achie ed (Figu e 9,∆𝐾=(𝐾10%−𝐾20%)⁡𝐾20%
⁄). Howe e , wi h a
con en o lignin g ea e han 30% he kine ics e olu ion changes. The kine ics o he
cellulosic pa o sample con inues dec easing wi h he amoun o lignin bu he kine ics
ela ed o lignin dec eases oo. This beha iou could be caused by an inc emen in he
he mal deg ada ion esis ance o he sample due o he ac ha lignin, which is he
s onges componen agains he mal deg ada ion, would be mo e han 30%.

17
Table 4: HERE
Figu e 9: HERE
I is ema kable ha kine ic cons an s ela ed wi h hemicellulose deg ada ion do no
change hei alues ( o his eason hey a e no p esen in Figu e 9). This esul could
be explained by he ac ha in his s udy case he e we e no p e ious ac o s ha can
sol e i . So, i s in e ac ions wi h cellulose and lignin would be independen o he lignin
concen a ion. The ole o he empe a u e is he same as in pa s 4.2.1.1 and 4.2.1.3
and K8 is no ep esen ed because he e is no change in i s alue.
4.2.2.2. Iso he mal p ocess
Finally, he TGA o a sample o Ace Saccha um was s udied in iso he mal condi ions
a 150, 200, 250, 300 and 350 ºC. The adjus men needed a se o pa ame e s o each
empe a u e, when i is highe han 200ºC, and di e en om he pa ame e s used in
he non-iso he mal p ocess (¡E o ! No se encuen a el o igen de la e e encia.).
Thei a e age absolu e de ia ions we e: 0.71%, 0.70%, 0.39%, 0.63% and 2.68%
espec i ely. This disc epancy be ween he kine ics due o he ype o p ocess could be
caused by a p o ec i e in e ac ion be ween species. This means ha , in an iso he mal
mode, he decomposi ion is low up o a ce ain empe a u e is eached (250ºC) and
hemicellulose deg ada ion s a s enhancing he decomposi ion o cellulose and lignin.
Besides, he e is an enhanced in lignin deg ada ion 300 ºC, when cellulose would s a
i s deg ada ion. This idea could explain he d as ic change in he kine ics cons an s
shown in Figu e 10 a 250ºC and 300ºC.
Figu e 10: HERE
I can be seen in Figu e 10 ha he he mal deg ada ion a iso he mal condi ions
depends on empe a u e s ongly (as i was expec ed). In his case K5 and K8 a e no
p esen in he g aph, because hey did no change. This would be caused by he ac
ha he maximum ope a ional empe a u e (350ºC) is no high enough o b eak lignin
o cellulose cha .
4.3. Composi ion es ima ion
18
Once all he expe imen al da a ha e been adjus ed, he capabili y o he model o
es ima e he ini ial composi ion o he biomass was es ed. The sample used o y his
es ima ion was he TGA o non-ex ac ed g ape seeds.
The p edic ion implies an op imiza ion p oblem in which he di e ence be ween he
expe imen al and simula ed TGA mus be minimized changing he alues o he ini ial
composi ion ( 12 ). The p oblem was limi ed by he ollowing es ain s. I was assumed
ha he e is no ini ial cha coal in he sample and ha he ini ial composi ion o wa e ,
oil, hemicellulose, cellulose and lignin we e in he anges showed in he
Table 5. The amoun o ine compounds was ob ained by balance o he o al.
min
𝑚𝑗(∑|𝑀𝐸𝑥𝑝−𝑀|
𝑡=𝑡𝑓
𝑡=0 );⁡𝑚𝑗𝑚𝑖𝑛 <𝑚𝑗<𝑚𝑗𝑚𝑎𝑥;𝑗∈[1,𝑁]
( 12
)
Table 5: HERE
The calcula ed composi ion is no e y accu a e because in some compounds he
de ia ion is high, o example he maximum de ia ion o wa e was 60.5 % (Table 6).
Bu aking in o accoun ha he alues o he maximum and he minimum o each
componen we e s ablished in a gene al way, he p edic ion is good enough. In o de o
imp o e hese alues, mo e expe imen al composi ions would be needed o ix a be e
op imiza ion ange. I is in e es ing ha he calcula ed cellulose composi ion is close o
he expe imen al one han he hemicellulose composi ion. This esul could be caused
by he ac ha he oil apo iza ion and hemicellulose deg ada ion can appea bo h
be ween 250ºC and 300ºC.
Table 6: HERE
5. Conclusions
An au o-ca aly ic kine ic model has been de eloped o he mog a ime ic analysis wi h
an a e age absolu e de ia ion be ween he simula ion and he expe imen al da a lowe
han 7% in all he s udied cases. This model can simula e he beha iou o e y
di e en samples (seeds, g ape skins and ees) as iso he mal p ocess as non-
iso he mal p ocess p o iding he composi ion p o iles o hei indi idual componen s
19
oo. In addi ion, he model can ep oduce he e ec o he hea ing a e in he
decomposi ion using a non-A henius’ dependence wi h he empe a u e. Due o he
ac ha he kine ics pa ame e s change wi h he ype o biomass i is deduced ha he
s uc u e o biomass has a e y impo an ole in he mal deg ada ion. Also impo an is
he composi ion because some species can wo k as a shield ha a oids he
deg ada ion o he o he s un il hei clea age s a . Finally, a p elimina y composi ion
es ima ion we e done, s a ing om a TGA cu e and es ima ing he composi ion o he
biomass ma e ial. This p edic ion has an accep able accu acy especially o cellulose
and lignin (di e ences lowe han 7%). Howe e , he p edic ion o he essen ial oil is
ick and in o de o inc ease model ideli y, mo e expe imen s would be needed, which
would allow o s ablish be e limi s o he op imiza ion anges o he ini ial composi ion
and o imp o e he kine ics pa ame e s.
T
Table 7: HERE
Acknowledgemen s
The au ho s acknowledge he Spanish Economy and Compe i i eness Minis y, P ojec
Re e ence: ENE2012-33613 and he egional go e nmen (Jun a de Cas illa y León),
P ojec Re e ence: VA330U13 o unding. The au ho s would like o hank P o . Ped o
Fa dim and D . Kons an in Gabo om Åbo Akademi o hei help wi h he hyd o opic
lignin.
Nomencla u e
Ac onyms
C: Cellulose.
HC: Hemicellulose.
L: Lignin.
O: Oil.
TGA: The mog a ime ic analysis.
W: Wa e .
XC: Cha coal p oduced o m cellulose.
XHC: Cha coal p oduced o m hemicellulose.
XL: Cha coal p oduced o m lignin.
20
Subindex and supe index
EXP: Expe imen al da a o he TGA.
in: ine compounds.
TOTAL: To al simula ed TGA.
G eek le e s and symbols
𝛼𝑖: Ini ializa ion ac o , adim.
𝛽𝑖: Accele a ion ac o , adim.
𝐴𝑗−𝐹𝑗: An oine’s equa ion coe icien s o he compound “j”, adim.
𝑐: Co ec ion ac o o he kine ic in he decomposi ion a di e en hea ing a es o he
cellulose, adim.
𝐶𝑗: Concen a ion o “j” in he gas phase, kmol/m3.
𝐶𝑗∗: Equilib ium concen a ion o “j” in he in e phase be ween he liquid and he gas
phase, kmol/m3.
𝐸𝑎𝑖
𝑅: Ac i a ion ene gy o he eac ion “i”, K.
ℎ: Pa ial mass ans e coe icien be ween he liquid and he gas, kgj· m3
/min·m2·kmolj.
𝑘𝑜𝑖: P eexponen ial ac o o he eac ion “i”, min-1.
𝑘𝑖: Kine ic cons an o he eac ion “i”, min-1.
𝑀𝑒𝑥𝑝: Expe imen al mass ac ion o un eac ed biomass, gsample/gsample ini ial.
𝑀: Mass ac ion o un eac ed biomass, gsample/gsample ini ial.
𝑚𝑗𝑚𝑎𝑥: Maximun alue o mass ac ion o he compound “j” in he biomass, g/g.
𝑚𝑗𝑚𝑖𝑛: Minimum alue o mass ac ion o he compound “j” in he biomass, g/g.
𝑚𝑗: Mass ac ion o he compound “j” in he biomass, g/g.
𝑁: Numbe o compounds in he biomass, adim.
𝑛𝑖: o de o eac ion o he eac ion “i”, adim.
𝑛𝑙𝑖: Mass ans e o de , adim.
𝑁𝑟: Numbe o eac ions, adim.ç
𝑃𝑗∗: Vapou p essu e o he compound “j”, a m.
𝑟𝑖: Reac ion eloci y numbe “i”, g/min·g.
𝑟𝑗: Reac ion eloci y o decomposi ion o he componen “j” in he biomass, g/min·g.
𝑆: Exchange su ace be ween he liquid and he gas, m2.
: Ope a ing ime, min.
𝑇: Ope a ing empe a u e, K.
21
𝒙𝒊𝒆𝒙𝒑: Expe imen al biomass ac ion, gsample/gsample ini ial.
𝒙𝒊𝑺𝑰𝑴: Simula ed biomass ac ion, gsample/gsample ini ial.
Lis o igu es
Figu e 1: Schema o he biomass s uc u e.
Figu e 2: Reac ion pa hway in a he mal decomposi ion.
Figu e 3: Fi ing o he hemicellulose decomposi ion wi h a hea ing a e o 20 ºC/min.
W: Wa e . HC: Hemicellulose. XHC: Cha o hemicellulose.TOTAL: Simula ed TGA.
EXP: Expe imen al TGA.
Figu e 4: Simula ed di e en ial he mog aphy o he hemicellulose TGA. W: Wa e . HC:
Hemicellulose. XHC: Cha o hemicellulose. TOTAL: Simula ed DTG.
Figu e 5: Simula ed di e en ial he mog aphy o he lime TGA. W: Wa e . HC:
Hemicellulose. XHC: Cha o hemicellulose. TOTAL: Simula ed DTG. O: Oil. C:
Cellulose. L: Lignin. XC: Cha o cellulose. XL: Cha o lignin.
Figu e 6: Simula ed di e en ial he mog aphy and expe imen al di e en ial
he mog aphy. EXP: Expe imen al DTG. TOTAL: Simula ed DTG.
Figu e 7: Va ia ion o he kine ic cons an s be ween he ex ac ed seeds and he no
ex ac ed seeds. K1: kine ic cons an o hemicellulose deg ada ion o ola iles. K2:
kine ic cons an o cellulose deg ada ion o ola iles. K3: kine ic cons an o lignin
deg ada ion o ola iles. K4: kine ic cons an o lignin deg ada ion o cha . K5: kine ic
cons an o lignin cha deg ada ion o ola iles. K6: kine ic cons an o cellulose
deg ada ion o cha . K7: kine ic cons an o hemicellulose deg ada ion o cha .
Figu e 8: Va ia ion o he kine ic cons an an be ween he hyd olysed seeds and he
non- hyd olysed seeds. K1: kine ic cons an o hemicellulose deg ada ion o ola iles.
K2: kine ic cons an o cellulose deg ada ion o ola iles. K3: kine ic cons an o lignin
deg ada ion o ola iles. K4: kine ic cons an o lignin deg ada ion o cha . K5: kine ic
cons an o lignin cha deg ada ion o ola iles. K6: kine ic cons an o cellulose
deg ada ion o cha . K7: kine ic cons an o hemicellulose deg ada ion o cha .

22
Figu e 9: Va ia ion in pe cen age o he eac ion kine ics be ween he samples be ween
26% and 30% o lignin. K2: kine ic cons an o cellulose deg ada ion o ola iles. K3:
kine ic cons an o lignin deg ada ion o ola iles. K4: kine ic cons an o lignin
deg ada ion o cha . K5: kine ic cons an o lignin cha deg ada ion o ola iles. K6:
kine ic cons an o cellulose deg ada ion o cha .
Figu e 10: Kine ics cons an in each iso he mal p ocess.K1: kine ic cons an o
hemicellulose deg ada ion o ola iles. K2: kine ic cons an o cellulose deg ada ion o
ola iles. K3: kine ic cons an o lignin deg ada ion o ola iles. K4: kine ic cons an o
lignin deg ada ion o cha . K6: kine ic cons an o cellulose deg ada ion o cha . K7:
kine ic cons an o hemicellulose deg ada ion o cha . K9: kine ic cons an o
hemicellulose cha deg ada ion o ola iles.
Lis o ables
Table 1: Ini ial composi ion o he samples.
Table 2: An oine’s equa ion coe icien s o wa e (W) and oil (O).
Table 3: A e aged mass ans e pa ame e s o wa e (W) and oil (O).
Table 4: G oups o woody samples aking in o accoun i s lignin con en .
Table 5: Ini ial composi ion a ia ion anges o he composi ion es ima ion o no-
ex ac ed g ape seeds.
Table 6: Compa ison be ween he es ima ed and expe imen al composi ion o no-
ex ac ed g ape seeds.
Table 7: Kine ics pa ame e s i ed o all he samples.
Re e ences
23
[1] Cla k JH, Buda in V, Deswa e FEI, Ha dy JJE, Ke on FM, Hun AJ, e al.
G een chemis y and he bio e ine y: A pa ne ship o a sus ainable u u e.
G een Chemis y. 2006;8:853-60.
[2] Bozell JJ. Feeds ocks o he u u e - Bio e ine y p oduc ion o chemicals
om enewable ca bon. Clean - Soil, Ai , Wa e . 2008;36:641-7.
[3] Cheng S, Zhu S. Lignocellulosic eeds ock bio e ine y- he u u e o he
chemical and ene gy indus y. BioResou ces. 2009;4:456-7.
[4] Ca ie M, Loppine -Se ani A, Denux D, Lasnie J-M, Ham-Picha an F,
Cansell F, e al. The mog a ime ic analysis as a new me hod o de e mine he
lignocellulosic composi ion o biomass. Biomass and Bioene gy. 2011;35:298-
307.
[5] Williams PT, Besle S. The In luence o Tempe a u e and Hea ing Ra e on
he Slow Py olysis o Biomass. Renewable Ene gy. 1996;7:233-50.
[6] L D, Xu M, Liu X, Zhan Z, Li Z, Yao H. E ec o cellulose, lignin, alkali and
alkaline ea h me allic species on biomass py olysis and gasi ica ion. Fuel
P ocessing Technology. 2010;91:903-9.
[7] Chen Q, Zhou JS, Liu BJ, Mei QF, Luo ZY. In luence o o e ac ion
p e ea men on biomass gasi ica ion echnology. Chinese Science Bulle in.
2011;56:1449-56.
[8] S. Völke TR. The mokine ic in es iga ion o cellulose py olysis. Impac o
ini ial and inal mass o kine ics esul s. Jou nal o Analy ical and Applied
Py olysis. 2002;62:165-77.
[9] Capa R, Khezami L, Bu nham AK. Assessmen o a ious kine ic models
o he py olysis o a mic og anula cellulose. The mochimica Ac a.
2004;417:79-89.
[10] Mangu V, Sabio E, Gañán J, González JF, Rami o A, González CM, e al.
The mog a ime ic s udy o he py olysis o biomass esidues om oma o
p ocessing indus y. Fuel P ocessing Technology. 2006;87:109-15.
[11] Zabanio ou A, Ioannidou O, An onakou E, Lappas A. Expe imen al s udy o
py olysis o po en ial ene gy, hyd ogen and ca bon ma e ial p oduc ion om
lignocellulosic biomass. In e na ional Jou nal o Hyd ogen Ene gy.
2008;33:2433-44.
[12] Slopiecka K, Ba occi P, Fan ozzi F. The mog a ime ic analysis and kine ic
s udy o popla wood py olysis. Applied Ene gy. 2012;97:491-7.
[13] Kas anaki E, Vam uka D, G ammelis P, Kaka as E. The mog a ime ic
s udies o he beha io o ligni e–biomass blends du ing de ola iliza ion. Fuel
P ocessing Technology. 2002;77–78:159-66.
[14] Elyounssi K, Colla d FX, Ma eke JAN, Blin J. Imp o emen o cha coal yield
by wo-s ep py olysis on eucalyp us wood: A he mog a ime ic s udy. Fuel.
2012;96:161-7.
[15] Van de Velden M, Baeyens, J., B ems, A., Janssens, B., Dewil, R.
Fundamen als, kine ics and endo he mici y o he biomass py olysis eac ion.
Renewable Ene gy. 2010;35:232-42.
[16] Meng LY, Kang SM, Zhang XM, Wu YY, Xu F, Sun RC. F ac ional
p e ea men o hyb id popla o accele a ed enzyma ic hyd olysis:
Cha ac e iza ion o cellulose-en iched ac ion. Bio esou ce Technology.
2012;110:308-13
[17] BOBLETER O. Hyd o he mal deg ada ion o polyme s de i ed om plan s.
P og Polym Sci. 1994;19:797-841.
24
[18] P ess W, Teukolsky S, Ve e ling W, Flanne y B. Nume ical ecipes 3 d
edi ion: The a o scien i ic compu ing. 2007.
25
Figu e 1: Schema o he biomass s uc u e.
CELULLOSE
HEMICELLULOSE
LIGNING
2
VOLATILES
6
CHARCELULLOSE
1
VOLATILES
7
CHARHEMICELULLOSE
3
VOLATILES
4
CHARLIGNIN
8
VOLATILES
9
VOLATILES
5
VOLATILES
Figu e 2: Reac ion pa hway in a he mal decomposi ion.
32
Holm oak
1.82
>30
Ceda
1.52
Ace
Saccha um
2.14
Almond
4.31
a A e age absolu e de ia ion be ween expe imen al and simula ed da a.
Table 8: Ini ial composi ion a ia ion anges o he composi ion es ima ion o no-
ex ac ed g ape seeds.
mmina
(g/g)
mmaxb
(g/g)
Wc
0.00
0.08
Od
0.00
0.20
HCe
0.10
0.25
C
0.15
0.60
Lg
0.15
0.45
a The lowes mass ac ion in he op imiza ion.
b The highes mass ac ion in he
op imiza ion. c Wa e con en . d Oil con en . e
Hemicellulose con en . cellulose con en . g
lignin con en .
Table 9: Compa ison be ween he es ima ed and expe imen al composi ion o no-
ex ac ed g ape seeds.
Wa
Ob
HCc
Cd
Le
in
m (w %)g
Expe imen al
0.0292
0.1655
0.1461
0.2142
0.4187
0.0263
Es ima ed
0.0469
0.1149
0.1887
0.2215
0.4121
0.0159
De ia ion (%)h
60.5%
-30.6%
29.2%
3.39%
-1.57%
-39.5%
a Wa e . b Oil. c Hemicellulose. d Cellulose. e Lignin. Ine . g Biomass composi ion in weigh pe cen age. h De ia ion
be ween he es ima ed and eal composi ion.
1

Table 7: Kine ics pa ame e s i ed o all he samples
2
3
4
5