J. Anal. Appl. Pyrolysis 83 (2008) 197–204
Contents lists available at ScienceDirect
Journal of Analytical and Applied Pyrolysis
journal homepage: www.elsevier.com/locate/jaap
Phosphorus catalysis in the pyrolysis behaviour of biomass
Daniel J. Nowakowski 1, Charles R. Woodbridge, Jenny M. Jones *
Energy and Resources Research Institute, School of Process, Environmental and Materials Engineering (SPEME), University of Leeds, Leeds, LS2 9JT, UK
A R T I C L E I N F O
A B S T R A C T
Article history:
Received 12 June 2008
Accepted 7 August 2008
Available online 19 August 2008
Phosphorus is a key plant nutrient and as such, is incorporated into growing biomass in small amounts.
This paper examines the influence of phosphorus, present in either acid (H3PO4) or salt ((NH4)3PO4) form,
on the pyrolysis behaviour of both Miscanthus giganteus, and its cell wall components, cellulose,
hemicellulose (xylan) and lignin (Organosolv). Pyrolysis–gas chromatography–mass spectrometry (PY–
GC–MS) is used to examine the pyrolysis products during thermal degradation, and thermogravimetric
analysis (TGA) is used to examine the distribution of char and volatiles. Phosphorus salts are seen to
catalyse the pyrolysis and modify the yields of products, resulting in a large increase in char yield for all
samples, but particularly for cellulose and Miscanthus. The thermal degradation processes of cellulose,
xylan and Miscanthus samples occur in one step and the main pyrolysis step is shifted to lower
temperature in the presence of phosphorus. A small impact of phosphorus was observed in the case of
lignin char yields and the types of pyrolysis decomposition products produced. Levoglucosan is a major
component produced in fast pyrolysis of cellulose. Furfural and levoglucosenone become more dominant
products upon P-impregnation pointing to new rearrangement and dehydration routes. The P-catalysed
xylan decomposition route leads to a much simpler mixture of products, which are dominated by furfural,
3-methyl-2-cyclopenten-1-one and one other unconfirmed product, possibly 3,4-dihydro-2-methoxy2H-pyran or 4-hydroxy-5,6-dihydro-(2H)-pyran-2-one. Phosphorus-catalysed lignin decomposition also
leads to a modified mixture of tar components and desaspidinol as well as other higher molecular weight
component become more dominant relative to the methoxyphenyl phenols, dimethoxy phenols and
triethoxy benzene. Comparison of the results for Miscanthus lead to the conclusion that the
understanding of the fast pyrolysis of biomass can, for the most part, be gained through the study of
the individual cell wall components, provided consideration is given to the presence of catalytic
components such as phosphorus.
ß 2008 Elsevier B.V. All rights reserved.
Keywords:
Pyrolysis
Miscanthus
Lignocellulose
Phosphorus
Catalysis
1. Introduction
Phosphorus is one of the key plant nutrients, and as such, has
variable concentrations in biomass and energy crops [1]. It is an
interesting element in fuels, since, like potassium, it influences not
only the thermal behaviour [2] but also the ash behaviour is
combustion systems [3]. A typical concentration of P2O5 in the ash of
willow is 11.5% [4], while values are lower for the grasses, of the
order of 2–4% [4,5]. A value of 5.3% P2O5 in the ash has been reported
previously for Miscanthus giganteus [4]. The mobilization and ash
behaviour of phosphorus during combustion is a topic generating
* Corresponding author. Tel.: +44 113 343 2744; fax: +44 113 246 7310.
E-mail addresses: d.j.nowakowski@aston.ac.uk (D.J. Nowakowski),
j.m.jones@leeds.ac.uk (J.M. Jones).
1
Current Address: Bioenergy Research Group, School of Engineering and Applied
Science, Aston University, Birmingham, B47ET, UK.
0165-2370/$ – see front matter ß 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.jaap.2008.08.003
interest since it impacts not only on slagging, corrosion and
emissions, but also on sustainability and the possible beneficial use
of ash residues [6,7].
Phosphorus compounds are well-known flame-retardants,
and increase char yields from textiles and woods [8,9]. They also
catalyse dehydration reactions of cellulose and a recent study by
Di Blasi et al. [10] report decreasing yields of tar products from
phosphorus impregnated fir wood. There has been some previous
work examining the mechanism of phosphoric acid catalysed
decomposition of biomass, particularly cellulose [11,12]. Dobele
et al. [11] used analytical pyrolysis combined with gas
chromatography to study the composition of volatile products
of different celluloses impregnated with various amounts of
phosphoric acid. The influence on the yields of levoglucosan and
levoglucosenone was studied taking into account the supramolecular structure, degree of polymerization, hydrophilic properties and pre-treatment conditions of the celluloses. It was found
that levoglucosenone predominates in the volatiles of acid
198
D.J. Nowakowski et al. / J. Anal. Appl. Pyrolysis 83 (2008) 197–204
catalysed pyrolysis. However, the relative total amount of both 1,6anhydrosaccharides varied only in a narrow range of 75–85%
regardless of the impregnation and pre-treatment conditions of the
celluloses. For pulps with a less ordered cellulosic supramolecular
structure, breaking of glycosidic bonds with the formation of
levoglucosan besides levoglucosenone occurred and the relative
amount of non-dehydrated anhydrosaccharides increased. Dobele
et al. [12] also studied the influence of phosphoric acid pretreatment of biomass materials (beech wood sawdust, recycled Kraft
pulp, newsprint, microcrystalline cellulose Thermocell) on the
pyrolysis products yields. Birch wood treated with 1% phosphoric
acid yields approximately 15% of levoglucosan and 8% levoglucosenone. At higher concentrations of phosphoric acid (2.5%) the
formation of more levoglucosenone (17%) was observed. These
authors presented a mechanism of acid catalysed splitting of
glycosidic bonds in cellulose, showing that the interaction of mineral
acids proceeds via protonation of the oxygen atom on glycosidic
bonds, stabilisation of the pyranosyl cation by mesomerism, the
formation of oxonium ions by water addition, and stabilisation of the
hydrogen splitting.
The aims of the present work are to examine the influence of
both phosphoric acid and ammonium phosphate on the products
from pyrolysis of other cell wall components, hemicellulose
(xylan) and lignin (Organosolv). Results are compared with the
analogous cellulose decomposition routes. A second aim is to
examine how well the pyrolysis of an energy crop, Miscanthus giganteus, can be described in terms of the pyrolysis
of its cell wall components.
2. Experimental
2.1. Materials
The following cell wall components: cellulose, hemicellulose
(oat spelt xylan) and lignin (Organosolv) were used in this study.
All compounds were purchased from the Sigma–Aldrich Company Ltd. The biomass sample of Miscanthus giganteus was
obtained from Rothamsted Research (Harpenden, Hertfordshire,
UK). The sample was ground and sieved. The fraction 0.15–
0.18 mm was used for demineralisation, impregnations and
analyses. The cell wall components of Miscanthus giganteus
were determined using a combination of methods from the
literature [13–15]. The biomass sample was first extracted by
soxhlet extraction with a mixture of 95% ethanol and toluene
(1:2 v/v) for 6 h followed by 95% ethanol for 4 h and then distilled
water for2 h. This yielded the extractives content on the basis of
the sample weight loss. Lignin content was determined as Klason
lignin, calculated as the weight loss after 72% sulphuric acid
treatment of the extractive free sample. Holocellulose samples
were prepared by delignifaction using acid chlorite, where
sodium chlorite was used for the reaction in acetate buffer
(pH = 3.5). Hemicellulose was removed by alkali extraction with
17.5% NaOH solution. The analysis was as follows on a dry ash
free basis: Extractives, 5.11%; Klason lignin, 17.34%; Cellulose,
51.26%; Hemicellulose, 26.29%.
2.2. Sample preparation
2.2.1. Demineralisation
Hydrochloric acid treatment of cellulose and lignin sample was
performed by heating of 10 g of sample in 50 cm3 of 2.0 M HCl for
6 h at 333 K. After 48 h the sample, left in the HCl solution, was
again heated at 333 K for 6 h. The sample was filtered, then washed
using de-ionized water until the filtrate was Cl free (checked by
0.1 M silver nitrate solution). The sample was then oven dried at
333 K to constant weight. The same procedure was applied for
Miscanthus biomass sample.
2.2.2. Impregnation
2.2.2.1. Ortho-phosphoric acid. H3PO4 – impregnation: 0.5 g of
sample (demineralised cellulose, lignin and Miscanthus as well
as raw xylan from oat spelt), analysed before impregnation for
moisture content, was impregnated by phosphorus to yield a
2 wt.% P-impregnated sample. 0.1 M solution of ortho-phosphoric
acid was used for impregnation. After addition of phosphorus
acetate the sample was moistened by approximately 5 cm3 of deionized water, mixed and then oven dried at 333 K to constant
weight. Please note: H3PO4 impregnated samples in text are called
‘acid impregnated’ samples.
2.2.2.2. Ammonium phosphate. (NH4)3PO4 – impregnation: The
following procedure was applied: 0.5 g of sample was added to
5 cm3 0.1 M H3PO4 solution and mixed (magnetic stirrer). The pH
of mixture was monitored by a pH-meter calibrated for two
calibration points (pH = 1 and pH = 7). The mixture was neutralised
(to pH = 7) by addition (dropwise) of 0.1 M solution of ammonium
hydroxide NH4OH. The mixture was mixed for further 5 min and
then oven dried at 333 K to constant weight. Please note: (NH4)3PO4
impregnated samples in text are called ‘neutralised’ samples.
2.3. Thermogravimetric analysis (TGA)
Pyrolysis tests were performed using a TGA analyser (Station
Redcroft Simultaneous Analyser STA-780 Series). A typical sample
mass of 10 mg was heated at 25 K/min in a purge of nitrogen with
the final temperature of 1173 K, and then the sample was held at
1173 K for 15 min.
2.4. Analytical pyrolysis tests (PY–GC–MS)
Pyrolysis–gas chromatography–mass spectrometry (PY–GC–
MS) tests were performed on each sample using a CDS 1000
pyrolyser coupled to an Agilent 5975 Series GC-MSD gas
chromatograph. The column was a RTX 1701 (14% cyanopropylphenyl, 86% dimethylpolysiloxane; 60 m, 0.25 mm i.d., 0.25 mm
d.f.). The gas chromatograph oven was held at 313 K for 2 min and
then programmed at 4 K/min to 523 K, held for 30 min. Approximately 2 mg of sample was placed in 20 mm quartz tube in
between quartz wool. The sample was pyrolysed at a set point
temperature of 873 K at a ramp rate of 20 K/ms with the final dwell
time of 20 s.
3. Results and discussion
3.1. Biomass components (model compounds study)
3.1.1. TGA pyrolysis
The differential thermogravimetric (DTG) results comparing
the impact of phosphorus in the TGA pyrolysis experiments of
cellulose, xylan and Organosolv lignin are shown in Fig. 1. Pyrolysis
yields from these studies are given in Table 1.
The decomposition of untreated cellulose (Fig. 1a) occurs in the
temperature region between 448 K and 676 K, with the maximum
peak temperature at 642 K. Demineralisation shifts the maximum
peak temperature to 631 K, within the same temperature region.
As discussed previously [2], this is thought to be due either to the
removal of catalytic metals through demineralisation, or due to a
modification (lowering) of the average cellulose MW by the
acid treatment. Phosphorus addition significantly changes the
D.J. Nowakowski et al. / J. Anal. Appl. Pyrolysis 83 (2008) 197–204
199
Table 1
Pyrolysis yields from TGA studies of treated model compounds
Sample
Pyrolysis yields (%)
Volatiles
Char
Cellulose raw
Cellulose HCl treated
Cellulose P-impregnated (acid)
Cellulose P-impregnated (neutralised)
Xylan raw
Xylan P-impregnated (acid)
Xylan P-impregnated (neutralised)
Lignin Organosolv raw
Lignin Organosolv HCl treated
Lignin Organosolv P-impregnated (acid)
Lignin Organosolv P-impregnated (neutralised)
92.9
93.1
77.6
69.6
76.9
66.2
68.7
60.5
59.1
51.7
53.1
7.1
6.9
22.4
30.4
23.1
33.8
31.3
39.5
40.9
48.3
46.9
and 30.4% for acid and neutralised sample, respectively. Increasing
char yield is thought to be the result of acid catalysed condensation
reactions [11].
The DTG profile for xylan pyrolysis is shown in Fig. 1b. The DTG
curve for the unimpregnated sample showed the peak maximum
at 582 K. The presence of phosphorus shifts the maximum peak
temperature downwards (by approximately 60 K) to 521 K for
the acid impregnated sample and 527 K for the neutralised
sample. The shape of DTG curve for raw xylan indicates that the
decomposition, once initiated, occurs more rapidly then for
the phosphorus impregnated samples. Phosphorus also increases
the char yield of xylan from 23.1% to 33.8% for the acid
impregnated sample and from 23.1% to 31.3% for the neutralised
sample (Table 1). We can speculate that both the catalytic action
and the increased char yields are brought about in an analogous
way to the cellulose sample, i.e. phosphorus salts catalyse
dehydratation reactions of xylans.
As shown in Fig. 1c the major mass loss for all lignin samples
occurs between 480 K and 800 K, giving broad peaks with the
maximum temperatures at 653 K and 661 K for raw and
demineralised sample, respectively. When phosphorus is added
to the lignin samples the maximum peak temperature of catalysed
pyrolysis is shifted slightly to higher temperature. Peak temperatures were observed at 686 K and 670 K for phosphoric acid
impregnated lignin and (NH4)3PO4 impregnated (neutralised)
lignin, respectively. The demineralisation process has no influence
on char yield, but for phosphorus impregnated samples the char
yields are increased by approximately 6–7% to values of 48.3% and
46.9 for acid and neutralised samples, respectively (Table 1).
Fig. 1. Differential thermogravimetric analysis (DTG) for the pyrolysis of differently
treated (a) cellulose, (b) xylan and (c) lignin.
decomposition profile. The acid impregnated cellulose sample
decomposes between 445 K and 572 K, with the maximum peak
temperature at 538 K. Ammonium phosphate impregnated cellulose (neutralised sample) has a very similar profile and starts to
decompose at 453 K with the peak temperature at 557 K; the
decomposition process is finished by 583 K. In both cases, singlestep decomposition processes are observed. The addition of
phosphorus to cellulose also has a dramatic influence on the
pyrolysis product distribution—significant increases in yields of
char are noted, from 6.9% for the demineralised sample to 22.4%
3.1.2. PY–GC–MS studies
Pyrolysis–gas chromatography–mass spectrometry (PY–GC–
MS) analysis has been introduced to study the generation of light
and medium volatile organics formed during pyrolysis. Assignments of the main peaks were made from mass spectral detection
(NIST05a MS library) and from the literature [16–18] and are given
under the each figure. Thus, the assignments given in this paper are
likely assignments based on mass spectra and previous work, but
have not been definitively elucidated. Chromatograms from PY–
GC–MS analyses of raw and demineralised cellulose were very
similar, and the trace for the latter is given in Fig. 2. Fig. 3 displays
the acid and neutralised phosphorus impregnated cellulose
samples. Selected cellulose key markers were identified for the
most abundant thermal degradation compounds, and their
structural formulas are added to Fig. 2.
The main cellulose thermal degradation products for
the uncatalysed pyrolysis have been analysed previously [19]
and, comprise: 2(5H)-furanone; furfural; 5-methyl-2-furancarboxaldehyde; 5-hydroxymethyl-2-furancarboxaldehyde; Sugars
200
D.J. Nowakowski et al. / J. Anal. Appl. Pyrolysis 83 (2008) 197–204
Fig. 2. Pyrolysis–GC–MS of demineralised cellulose.
The main peaks are assigned from mass spectral detection as follows: 1: furan; 2: 3methyl-furan; 3: 2-propenoic acid, methyl ester; 4: 2(5H)furanone; 5: furfural; 6:
2-propyl-furan; 7: 1-(2-furanyl)-ethanone; 8: 1,2-cyclopentanedione; 9: 5-methyl2-furancarboxaldehyde; 10: 3-methyl-1,2-cyclopentanedione; 11: 2,4-dihydroxy6-methyl-2H-pyran-2-one; 12: 3-furan-carboxylic acid, methylester; 13:
levoglucosenone; 14: 3,5-dihydroxy-2-methyl-4H-pyran-4-one; 15: 3-methyl1,2-cyclopentanediol; 16: 1,4:3,6-dianhydro-a-D-glucopyranose; 17: Unknown;
18: 5-hydroxymethyl-2-furancarboxaldehyde; 19: 1,2-cyclohexanedione; 20: Not
confirmed; 21: D-allose; 22: levoglucosan.
Fig. 3. Pyrolysis–GC–MS of (a) H3PO4 and (b) (NH4)3PO4 impregnated cellulose.
The main peaks are assigned from mass spectral detection as follows: 1: furan; 2: 2methyl-furan; 3: 2-butanone; 4. acetic acid; 5: 2(5H) furanone; 6: furfural; 7: 2propylfuran; 8: 5-methyl-2(3H) furanone; 9: 1-(2-furanyl)-ethanone; 10: 2cyclopenten-1,4-dione; 11: 5-methyl-2-furancarboxaldehyde; 12: 5-methyl-2(5H)furanone; 13: 3-methyl-1,2-cyclopentanedione; 14: phenol; 15: 3-furancarboxylic acid, methylester; 16: levoglucosenone ; 17: 1,4:3,6-dianhydro-a-Dglucopyranose; 18: 5-hydroxymethyl-2-furancarboxaldehyde; 19: 4-O-b-Dgalactopyranosyl-a-D-glucopyranose;
20:
Unknown
21:
D-allose;
22:
levoglucosan; 23: Unknown.
Fig. 4. Peak areas for key sugars from PY–GC–MS analysis of cellulose samples. (Peak
areas have been normalised per mg of volatile products.)
derivatives: (i) 1,6-anhydro-3,4-dideoxy-D3-b-D-pyranosen-2one (levoglucosenone); (ii) 1,4:3,6-dianhydro-a-D-glucopyranose;
(iii) 4-O-b-D-galactopyranosyl-a-D-glucopyranose; (iv) D-allose
and (v) 1,6-anhydro-b-D-glucopyranose (levoglucosan). Levoglucosan is reported as the major pyrolysis product of cellulose and is
formed under neutral [20] or acid [21] conditions and it is generally
accepted that generation of this anhydrosugar is the first step of
the formation of other volatile compounds. This work has
identified the other sugar derivatives listed above in addition to
levoglucosan, although it is unclear if they are parallel (alternative)
reaction products, or rearrangement from levoglucosan.
The presence of phosphorus changes decomposition profile.
The major volatile products are: furfural; 5-methyl-2-furancarboxaldehyde, levoglucosenone, 1,4:3,6-dianhydro-a-D-glucopyranose; 5-hydroxymethyl-2-furancarbox-aldehyde and D-allose.
Levoglucosan is still produced in similar yield, but it is no longer
the major component. No large differences between the profiles
(i.e. peak intensity/abundance) of acid and neutralised Pimpregnated cellulose sample were observed.
Fig. 5. Pyrolysis–GC–MS chromatogram for raw xylan.
The main peaks are assigned from mass spectral detection as follows: 1: 3-methyl1,2-cyclopentanedione; 2: acetone; 3: acetic acid; 4: propanioc acid; 5: acetic acid
anhydride with formic acid; 6: 2,3-pentanedione; 7: 2,3-dihydro-1,4-dioxin; 8: 3methyl butanal; 9: 1-hydroxy-2-butanone; 10: butanediol; 11: furfural; 12: 2methyl-2-cyclopenten-1-one; 13: 1,2-cyclopentanedione; 14: 3-methyl-2cyclopenten-1-one; 15: Unknown; 16: 3-methyl-1,2-cyclopentanedione; 17: 1,4dimethyl-1,3-cyclopentanedione; 18: phenol; 19: 2-methoxyphenol; 20: 2,3dihydroxybenzaldehyde; 21: 3-methylphenol; 22: 2,5-dimethyl phenol; 23:
cyclohexane; 24: sucrose; 25: 2-methoxy-4-vinyl phenol; 26:1,2-benzenediol;
27: 2,6-dimethoxy phenol; 28: 3-methyl-1,2-benzenediol; 29: D-mannose; 30:
hydroquinone; 31: 3 hydroxybenzaldehyde; 32: 2-methyl-,1,4-benzenediol; 33:
30 ,50 - dimethoxyacetophenone; 34: 3,4-dihydro-6-hydroxy-2H-1-benzopyran; 35:
2,6-dimethoxy-4-(2-propenyl)-phenol.
D.J. Nowakowski et al. / J. Anal. Appl. Pyrolysis 83 (2008) 197–204
201
Figs. 5 and 6, respectively. The main decomposition products of the
raw xylan sample are: propanoic acid; 1-hydroxy-2-butanone;
furfural; 3-methyl-1,2-cyclopentanedione; 2-methoxy-4-vinyl
phenol; 1,2-benzenediol and 2-methyl-1,4-benzenediol. The
influence of added phosphorus on the nature of the xylan
decomposition products is high, and produces a very noticeable
change in the product distribution. Main degradation products are
furfural (peak 5, retention time 14.2 min) and 3-methyl-2cyclopenten-1-one (peak 8, retention time 19.8 min). A third
main decomposition product (peak 9, retention time 21.1 min) is
tentatively assigned 3,4-dihydro-2-methoxy-2H-pyran with m/z
114 (parent ion?) and a strong fragmentation signal at m/z 58; or 4hydroxy-5,6-dihydro-(2H)-pyran-2-one as cited by Fahmi et al.
during pyrolysis of Lolium and Festuca grasses [23]. However, the
NIST05a library probability match was low, suggesting co-elution.
This simplified product distribution was present for both acid and
neutralised phosphorus impregnated xylan.
The chromatograms for the HCl treated lignin (Organosolv) and
H3PO4 impregnated lignin are shown in Fig. 7. No difference
between the product distributions from raw and HCl treated lignin
samples as well as between H3PO4 and (NH4)3PO4 impregnated
samples were observed. Thus, the comparison has been made
between the HCl treated sample and the H3PO4 impregnated
samples. Although the relative amounts change, the main
Fig. 6. Pyrolysis–GC–MS of (a) H3PO4 and (b) (NH4)3PO4 impregnated xylan
samples.
The main peaks are assigned from mass spectral detection as follows: 1:acetone; 2:
2-methyl-furan; 3: acetic acid; 4: 2,3-pentanedione; 5: furfural; 6: 1,2cyclopentanedione; 7: 5-methyl-2-3H-furanone; 8: 3 methyl-2-cyclopenten-1one; 9: 3,4-dihydro-2-methoxy-2H-pyran (or) 4-hydroxy-5,6-dihydro-(2H)-pyran2-one; 10: 3-methyl-1,2 cyclopentanedione; 11: 2 methoxyphenol; 12: 3methylphenol; 13: 2,5- dimethyl phenol; 14: 12-methoxy-4-vinyl phenol; 15:
1,2 benzenediol; 16: 3-methyl-1,2-benzenediol; 17: 3 hydroxybenzaldehyde; 18:
1-(3-hydroxyphenyl)-ethanone; 19: 3,4-dihydro-6-hydroxy-2H-1-benzopyran;
20: 2,6-dimethoxy-4-(2-propenyl)-phenol.
Peak area percentages for sugar derivatives from the chromatograms in Figs. 2 and 3 were normalised per mg of volatile products.
In this way peak areas for key sugar derivatives from the PY–GC–
MS profiles could be compared in terms of the catalytic impact of
phosphorus. This comparison is given in Fig. 4. (Furfural, a sugar
decomposition product, is a major component in phosphoruscatalysed samples but not included in Fig. 4.) In both cases (H3PO4
and (NH4)3PO4 impregnation), there is a significant influence of
phosphorus on the formation of levoglucosenone and furfural and
these are the main components. It is worth noting that the peak
area percentages for the levoglucosenone were increased from the
value of 5.2% to 51.0% (acid impregnation) and 30.9% (ammonium
phosphate impregnation). This indicates strong catalytic impact of
phosphorus on cellulose decomposition, favouring levoglucosenone formation as reported previously by Dobele et al. [12] and
Di Blasi et al. [10]. For our sample and conditions we see
levoglucosenone:levoglucosan ratios to increase markedly upon
addition of phosphorus, and this ratio is higher for the acid
impregnated sample, compared to the ammonium phosphate
impregnated sample. The highest levoglucosan yield for the
demineralised cellulose sample, as the major 1,6-anhydrosaccharide product for the uncatalysed reaction [22].
Chromatograms from PY–GC–MS analyses of xylan as well as
acid and neutralised P-impregnated xylan samples are given in
Fig. 7. Pyrolysis–GC–MS of (a) demineralised and (b) H3PO4 impregnated
Organosolv lignin.
The main peaks are assigned from mass spectral detection as follows: 1: phenol; 2: 2methoxy-phenol; 3: 2-methyl-phenol; 4: 2-methoxy-3 methyl-phenol; 5: 2methoxy- 4-methyl-phenol; 6: 3,5-dimethyl phenol; 7: 4-ethyl-phenol; 8: 4-ethyl
2-methoxyphenol; 9: 3-methoxy-1,2-benzenediol; 10: 1,2 benzenediol; 11: 2,6
dimethoxyphenol; 12: 3,4-dimethoxy-phenol; 13: 2-methoxy-4 (1-propyl)-phenol;
14: 1,2,4- triethoxybenzene; 15: vanillin; 16: 2,5-diethyl-hydroquinone; 17: 1 (2,3,4trihydroxymethyl)-ethanone; 18: 1,2,3-trimethoxy-5methyl benzene; 19: 1-(4hydroxy-3methoxyphenyl)-ethanone; 20: 2 methoxy-4-(methoxymethyl)-phenol;
21: 1-ethyl-3-(phenylmethyl)-benzene; 22: 3-(4-hydroxy-3-methoxyphenyl)-2propenoic acid; 23: 4-hydroxy-3-methoxy benzeneacetic acid; 24: 2,6 dimethoxy4-(2- propenyl)-phenol; 25: 4-hydroxy-3,5-dimethoxy benzaldehyde; 26: 3,5dimethoxy-4- hydroxyphenylacetic acid; 27: 1-(4 hydroxy-3,5-dimethoxyphenyl)
ethanone; 28: desaspidinol; 29: 1,2-dimethoxy-4-(1,2,3-methoxypropyl) benzene;
30: 30 ,50 -dimethoxyacetophenone; 31: aspidinol.
202
D.J. Nowakowski et al. / J. Anal. Appl. Pyrolysis 83 (2008) 197–204
decomposition products of demineralised and both acid and
neutralised lignin are: 2-methoxy-phenol; 2-methoxy-4-methylphenol; 3-methoxy1,2-benzenediol; 2,6-dimethoxy-pheno; 1,2,4trimethoxy-benzene; 1,2,3-trimethoxy-5-methyl-benzene; 2,6dimethoxy-4-(2-propenyl)-phenol; 1,2-dimethoxy-4-(1,2,3-methoxypropyl)-benzene, desaspidinol and aspidinol. However, the
presence of phosphorus during the pyrolytic decomposition of
the lignin matrix decreases the abundance of 1,2,4-trimethoxybenzene; 2-methoxy-4-methyl-phenol and 1,2,3-trimethoxy5-methyl-benzene. Vanillin and 2,5-diethyl-hydroquinone were
only detected in the raw and demineralised samples, while
phosphorus catalysis promotes the generation of 1-(4-hydroxy3methoxyphenyl)-ethanone; 3,5-dimethoxy-4-hydroxyphenylacetic acid and 30 ,50 -dimethoxyacetophenone (not present in raw and
HCl treated lignin samples). Phosphorus also significantly increased
the abundance of desaspidinol and 1,2-dimethoxy-4-(1,2,3-methoxypropyl)-benzene. Also the larger amount of unresolved material
eluting after 46 min is observed.
3.2. Miscanthus giganteus
3.2.1. TGA pyrolysis
Differential thermogravimetric results comparing the influence of phosphorus on the pyrolysis of the Miscanthus
giganteus biomass samples are shown in Fig. 8. The cell wall
components of Miscanthus are 51.26% cellulose, 26.29% hemicellulose, 17.34% lignin. Raw Miscanthus has two main unresolved peaks. The first step of thermal degradation is attributed to
the decomposition of hemicellulose and the initial stage of the
degradation of cellulose, while the second step is attributed to the
degradation of lignin and the final degradation of cellulose.
Comparison with the cell wall components studied individually in
Fig. 1, indicates that there is good agreement between peak
maximum temperatures (582 K vs. 596 K for the hemicellulose
decomposition peak, and 642 K vs. 649 K for the cellulose
decomposition peak). Upon acid treatment to remove inorganic
constituents from the biomass the first peak becomes much
weaker indicating some change in the hemicellulose content.
Also, the second main peak shifts to slightly higher temperature
(peak temperature 654 K), which shows that some catalytic
species have been removed by acid washing. In agreement with
the work presented on the cell wall components in Fig. 1, upon
addition of phosphorus, a strong catalytic effect on the degradation is observed. Peak maxima are shifted towards lower values:
546 K for acid impregnated Miscanthus and 554 K for neutralised
Fig. 8. DTG profiles for the pyrolysis of differently treated Miscanthus giganteus.
Table 2
Pyrolysis yields from TGA studies of treated Miscanthus samples
Sample
Miscanthus
Miscanthus
Miscanthus
Miscanthus
Pyrolysis yields (%)
raw
HCl treated
P-impregnated (acid)
P-impregnated (neutralised)
Volatiles
Char
81.9
90.4
68.1
71.5
18.1
9.6
31.9
28.5
sample (cf. 538 K and 557 K for acid and neutralised Pimpregnated cellulose, respectively).
The yields for TGA pyrolysis of Miscanthus samples are given in
Table 2. HCl treatment decreases the char yield (from 18.1% for raw
Miscanthus to 9.6% for the demineralised one), presumably
because inorganic constituents that promote char formation are
no longer present, a similar finding has been observed for willow
[2]. The presence of phosphorus is again seen to have a large
influence on the char formation stage, increasing the char yield,
and this is more significant in the acid impregnated sample. Thus, it
appears that phosphorus introduced as H3PO4 is more effective in
catalysing pyrolysis in the Miscanthus giganteus, as seen by a
larger decrease in the peak maximum temperature in the DTG plot
(Fig. 8), and a larger increase the char yield (Table 2).
3.2.2. PY–GC–MS studies
Chromatograms from PY–GC–MS analyses demineralised Miscanthus as well as acid and neutralised P-impregnated samples are
given in Figs. 9–11, respectively. There are some similarities
between the raw and demineralised Miscanthus giganteus
samples and only the latter is given in Fig. 9. The same cellulose
and lignin key markers were detected in pyrolysis products of both
samples. Peaks are identified for: cyclopentane derivatives (i.e. 1,2cyclopentanedione, 3-methyl-1,2-cyclopentanedione); 2(5H)furan; furfural; phenol and phenol derivatives (i.e 2-methoxyphenol, 3- and 4-methyl-phenol, 4-ethyl-phenol, 2-methoxy-4methyl-phenol, 2,6-dimethoxyphenol; 2-methoxy-4-(2-propenyl)phenol); 1,2,4-trimethoxybenzene; 30 ,50 -dimethoxyacetophenone; 1,2-benzenedimethanol.
Fig. 9. . Pyrolysis–GC–MS of demineralised Miscanthus.
The main peaks are assigned from mass spectral detection as follows: 1: 2(5H)furanone; 2: furfural; 3: 1,2-cyclopentanedione; 4: 5-methyl 2-furancarboxaldehyde;
5: 3-methyl 1,2-cyclopentanedione; 6: 2-methyl 1,2-cyclopentanedione; 7: phenol;
8: 2-methoxy-phenol; 9: 3- or/and 4 methyl-phenol; 10: levoglucosenone; 11: 2methoxy-4-methyl-phenol; 12: 4-ethyl-phenol; 13: 4-ethyl-2 methoxy-phenol; 14:
1,4:3,6 dianhydro-a-D-glucopyranose; 15: 2 methoxy-4-vinyl-phenol; 16: eugenol;
17: 1,2-benzenediol; 18: 2,6 dimethoxyphenol; 19: 4-methyl-1,2 benzenediol; 20: 2methoxy-4-(1 propenyl)phenol; 21: 1,2,4 trimethoxybenzene; 22: 2-methoxy 4propylphenol; 23: 30 ,50 dimethoxyacetophenone; 24: levoglucosan; 25: 2-methoxy4- (2-propenyl)phenol.
D.J. Nowakowski et al. / J. Anal. Appl. Pyrolysis 83 (2008) 197–204
Fig. 10. Pyrolysis–GC–MS for acid P-impregnated Miscanthus.
The main peaks are assigned from mass spectral detection as follows: 1: 2methylfuran; 2: 2-butanone; 3: acetic acid; 4: 2(5H)furanone; 5: furfural; 6: 1-(2furanyl)- ethanone; 7: 5-methyl-2-furancarboxaldehyde; 8: phenol; 9: 2-methoxyphenol; 10: furyl-hydroxymethyl ketone; 11: levoglucosenone; 12: 4-ethyl-phenol;
13: 1,4:3,6- dianhydro-a-D glucopyranose; 14: 2-methoxy-4 vinyl-phenol; 15: 5hydroxymethyl- 2-furan-carboxaldehyde; 16: 4 methyl-1,2-benzenediol; 17: 2
methoxy-4- propylphenol; 18: 1-(4-hydroxy-3-methoxyphenyl)-2-propanone; 19:
2,4’-dihydroxy- 30 -methoxyacetophenone; 20: D-allose; 21: levoglucosan; 22: 3,5
dimethoxy-4- hydroxyphenylacetic acid; 23: 1,6-anhydro-a-D galactofuranose; 24:
desaspidinol.
The main decomposition products of acid impregnated and
neutralised Miscanthus samples (Figs. 10 and 11) are: 2(5H)furan, furfural; 5-methyl-2-furancarboxaldehyde, phenol and
phenol derivatives (2-methoxy-phenol; 4-ethyl-phenol; 2-methoxy-4-propylphenol); 2-methoxy-4-vinyl-phenol and sugars
(dianhydroglucopyranose and levoglucosan). Levoglucosenone
is seen as an intense peak, which agrees with the findings from the
studies of P-impregnated cellulose (see Fig. 3). Furfural is also
detected as a high intensity product from pyrolysis of the Pimpregnated Miscanthus, and this was duplicated in pyrolysis
Fig. 11. Pyrolysis–GC–MS for neutralised P-impregnated Miscanthus.
The main peaks are assigned from mass spectral detection as follows: 1: 2methylfuran; 2: 2-butanone; 3: acetic acid; 4: 2(5H)furanone; 5: furfural; 6: 1-(2furanyl)- ethanone; 7: 5-methyl-2-furancarboxaldehyde; 8: phenol; 9: 2-methoxyphenol; 10: furyl-hydroxymethyl ketone; 11: levoglucosenone; 12: 4-ethyl-phenol;
13: 1,4:3,6-dianhydro-a-D glucopyranose; 14: 2-methoxy-4 vinyl-phenol; 15: 5hydroxymethyl-2-furan-carboxaldehyde; 16: 4-methyl 1,2-benzenediol; 17: 2methoxy-4- propylphenol; 18: 1-(4-hydroxy-3 methoxyphenyl)-2-propanone; 19:
2,4’-dihydroxy- 3’ methoxyacetophenone; 20: D-allose; 21: levoglucosan; 22:
desaspidinol.
203
Fig. 12. Peak areas for key cellulose and lignin markers from PY–GC–MS analysis for
Miscanthus samples (all treatments). (Peak areas have been normalised per mg of
volatile products.)
studies of both cellulose and xylan (Figs. 3 and 6). However, one of
the major products from P-impregnated xylan (peak 9, Fig. 6) is
not evident in high concentrations in the P-impregnated
Miscanthus (Fig. 9), even though Miscanthus contains 26%
hemicelluloses. However, there may be branching and crosslinking differences in the individual hemicellulose contents of
Miscanthus compared to oat spelt xylan, which would impact on
the nature of the pyrolysis fragments.
Some peak area percentages from the chromatograms in
Figs. 9–11 were normalised per mg of volatile products and
compared in Fig. 12. The peak areas for the three sugar derivatives
(levoglucosenone, 1,4:3,6-dianhydro-a-D-glucopyranose and
levoglucosan) are compared in Fig. 13. It is clear that phosphorus
has a significant influence on the formation of volatile products.
Acid (phosphorus) treatment promotes decomposition of levoglucosan and significantly increased the yield of levoglucosenone and
1,4:3,6-dianhydro-a-D-glucopyranose at the expense of levoglucosan. The amount of the furfural in the tars also increases.
Demineralisation lowers yields of methoxyl phenols and 2methoxy-4-vinyl-phenol. The latter is one of the major of the
products from raw Miscanthus and presumably arises from xylan
decomposition (cf. Fig. 5), since demineralisation removes some of
the hemicellulose component.
Fig. 13. Peak areas for key sugar derivatives from PY–GC–MS analysis for
Miscanthus samples (all treatments). (Peak areas have been normalised per mg of
volatile products.)
204
D.J. Nowakowski et al. / J. Anal. Appl. Pyrolysis 83 (2008) 197–204
4. Conclusions
The main cell wall constituents (cellulose, xylan, lignin) as well
as Miscanthus giganteus sample were subjected to three types of
pre-treatment – HCl demineralisation and impregnation of the
demineralised samples by ortho-phosphoric acid – H3PO4 – and
ammonium phosphate – (NH4)3PO4.
In this study it was observed that the phosphorus salts
catalysed the pyrolysis and that the yields of pyrolysis products
were modified. The phosphorus-catalysed pyrolytic decomposition resulted in a large increase in char yield for all samples, but
particularly for cellulose and Miscanthus. For example the char
yield increased from 6.9% for demineralised cellulose to 22.4% and
30.4% for the acid impregnated and neutralised samples. Also in
the case of Miscanthus sample char yield is seen to triple. DTG
analysis revealed that the thermal degradation processes of
cellulose, xylan and Miscanthus samples occur in one step and
the main pyrolysis step is shifted to lower temperature by
approximately 100 K. A small impact of phosphorus was observed
in the case of lignin pyrolysis char yields and types of pyrolysis
decomposition products produced.
The tar components produced from phosphorus impregnated
and demineralised samples were very different. Levoglucosan is a
major component produced in fast pyrolysis of cellulose. Furfural
and levoglucosenone become more dominant products upon Pimpregnation pointing to new rearrangement and dehydration
routes. This is true regardless of whether phosphorus is added as
phosphoric acid or ammonium phosphate.
The P-catalysed xylan decomposition route leads to a much
smaller mixture of products, which are dominated by furfural, 3methyl-2-cyclopenten-1-one and one other product, possibly 3,4dihydroxy-2-methoxy-2H-pyran or 4-hydroxy-5,6-dihydro-(2H)pyran-2-one.
Phosphorus-catalysed lignin decomposition route also leads to
a modified mixture of tar components and desaspidinol as well as
other higher molecular weight component become more dominant
relative to the methoxyphenyl phenols, dimethoxy phenols and
triethoxy benzene.
The results for Miscanthus corroborated the cell wall component work where, with one exception, similar products were
identified. This again leads to the conclusion that the under-
standing of the fast pyrolysis of biomass can be gained through the
study of the individual cell wall components, provided consideration is given to the presence of catalytic components such as
phosphorus (and potassium).
The work gives mechanistic insight into P-catalysed pyrolysis
and has implication in the production of useful chemicals and
products from biomass pyrolysis.
Acknowledgement
The authors are grateful to the EPSRC for the financial support of
this work through an Advanced Research Fellowship (JMJ) and
associated research grant (GR/S49018 and GR/S49025).
References
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
A. Monti, N. Di Virgilio, G. Venturi, Biomass Bioenergy 32 (2008) 216.
D.J. Nowakowski, J.M. Jones, R.M.D. Brydson, A.B. Ross, Fuel 86 (2007) 2389.
B.M. Jenkins, L.L. Baxter, T.R. Miles Jr., T.R. Miles, Fuel Proc. Technol. 54 (1998) 17.
D.G. Christian, A.B. Riche, Establishing fuel specifications of non-wood biomass
crops, ETSU B/U100612/REP 1999.
A. Moilanen, L.H. Sørensen, T.E. Gustafsson, J. Laatikainen-Luntama, E. Kurkela, in:
A.V. Bridgwater (Ed.), Progress in Thermochemical Biomass Conversion, IEA
Bioenergy, vol.1, Blackwell Science, Oxford, 2001, p. 122.
E. Lindstom, M. Sandstrom, D. Bostrom, M. Ohman, Energy Fuels 21 (2007) 710.
J. Beck, S. Unterberger, Fuel 85 (2006) 1541.
S. Gaan, G. Sun, J. Anal. Appl. Pyrolysis 78 (2007) 371.
R. Stevens, D.S. Van Es, R. Bezemer, A. Kranenbarg, Polym. Degrad. Stab. 91 (2006)
832.
C. Di Blasi, C. Branca, A. Galgano, Polym. Degrad. Stab. 93 (2008) 335.
G. Dobele, D. Meier, O. Faix, S. Radtke, G. Rossinskaja, G. Telysheva, J. Anal. Appl.
Pyrolysis 58–59 (2001) 453.
G. Dobele, T. Dizhbite, G. Rossinskaja, G. Telysheva, D. Meier, S. Radtke, O. Faix, J.
Anal. Appl. Pyrolysis 68–69 (2003) 197.
T. Ona, T. Sonoda, M. Shabata, K. Fukazawa, Tappi J. 78 (1995) 121.
E. Meszaros, E. Jakab, G. Varhegyi, J. Anal. Appl. Pyrolysis 79 (2007) 61.
J. Rodrigues, D. Meier, O. Faix, H. Pereira, J. Anal. Appl. Pyrolysis 48 (1999) 121.
O. Faix, I. Fortmann, J. Bremer, D. Meier, Holz Roh Werkst. 49 (1991) 213.
O. Faix, D. Meier, I. Fortmann, Holz Roh Werkst. 48 (1990) 281.
B.R.T. Simoneit, Appl. Geochem. 17 (3) (2002) 129.
D.J. Nowakowski, J.M. Jones, J. Anal. Appl. Pyrolysis 83 (2008) 12–25.
F. Safizadeh, Y.L. Fu, Carbohydr. Res. 29 (1) (1973) 113.
A. van Der Kaaden, J. Haverkamp, J.J. Boon, J.W. De Leeuw, J. Anal. Appl. Pyrolysis 5
(3) (1983) 199.
R.J. Evans, T.A. Milne, Energy Fuels 1 (1987) 123.
R. Fahmi, A.V. Bridgwater, S.C. Thain, I.S. Donnison, P.M. Morris, N. Yates, J. Anal.
Appl. Pyrolysis 80 (2007) 16.