Content uploaded by Daniele Duca
Author content
All content in this area was uploaded by Daniele Duca on Oct 17, 2018
Content may be subject to copyright.
1
The final publication is available at Springer via
http://dx.doi.org/10.1016/j.biombioe.2013.05.012
Published in BIOMASS AND BIOENERGY, vol. 56 September 2013, Pages 317-322
FINAL ACCEPTED MANUSCRIPT
Investigation on wood pellet quality and relationship between ash content and the most
important chemical elements
G. Toscano, G. Riva, E. Foppa Pedretti, F. Corinaldesi, C. Mengarelli, D. Duca
Dipartimento di Scienze Agrarie, Alimentari e Ambientali, Università Politecnica delle Marche. Via
Brecce Bianche – 60131 Ancona (AN) – Italy. Corresponding author e-mail: d.duca@univpm.it
tel. 011 39 071 2204297, fax. 011 39 071 2204167
Abstract
During the past four years, the European wood pellet market is currently continuing to increase.
Total European consumption was about 9.8 Tg and Italian market reaches almost 1 Tg mainly for
domestic supply. The present investigation provides an evaluation of pellet quality of Italian market
by analysis of 88 wood pellet samples acquired randomly and directly from the sale points.
Particular attention has been dedicated to relationship between ash and some elements. Results
highlight that only half of samples fulfill the A1 class (best quality) requirements established in the
EN 14961-2. Statistical analysis pointed out that ash content is mainly linked to sulphur, potassium
and chlorine. The results show that low limit threshold values for ash content ensure low chlorine
and sulfur contents.
Keywords: biomass characterization, EN 14961-2, standard, ash, chlorine, sulphur, correlation.
1. Introduction
During the recent years, wood pellet has become an important worldwide fuel. At global level wood
pellet industry production reaches about 13 Tg. Pellet has become popular in many countries,
especially in Europe, where its market is nowadays a large business, and is currently undergoing
rapid development. About 9.2 Tg were consumed by European countries in 2009 [1]. European
Biomass Association expects a consumption of 50 Tg for the European countries by 2020 [2-3].
Nowadays, the rapid growing of pellet market is so substantial that the wood pellet has to be
considered one of the largest internationally traded solid biomass commodities [4]. Furthermore,
even though most pellets in the past were made from pure wood (sawdust or shavings), the increase
of pellet demand entailed the supply of pure wood to be insufficient and other raw materials were
used. Currently bark, branches and stem wood have come into use. The variability of characteristics
of these materials leads to a greater need of defining qualitative standards, which is an indispensable
requirement for orienting the market [5-6].
2
Regarding the Italian situation, despite being the third largest European pellet producer, its
consumption in 2009 totaled over 1 Tg exceeding significantly the domestic supply [1]. In Italy
wood pellets are predominantly used in residential applications such as boilers and stoves [7],
therefore product quality is particularly important.
Just to ensure the quality of the pellets, European standards for solid biofuels are continuously
under development by CEN technical committee 335. The main European standard for wood pellet
is the EN 14961-2 fuel specification which defines several levels of product quality of wood pellets
for non-industrial use [8-9]. This standard sets out ranges for several parameters that are relevant to
assess wood pellet quality (and other solid biofuels quality) and indicates the technical standards
necessary for laboratory measurements. One of the most important parameter refers to ash content
(inorganic materials) representing a serious problem in biomass combustion because it causes
slagging, bed agglomeration, fouling, and corrosion in the combustion device, which degrades its
performance and severely damages the firing equipment [10]. Moreover, ash in wood biomass
consists of Cl, S, major elements (Al, Si, K, Na, Ca, Mg, Fe, P and Ti) and minor elements (As, Ba,
Be, Cd, Co, Cr, Cu, Hg, Mo, Mn, Ni, Pb, Se, Te, V and Zn) which directly affect aerosol and fly ash
formation during wood fuel combustion [11]. Other parameters like durability and bulk density are
also important for pellet storage and handling. Pellets are susceptible to mechanical wear, which
leads to production of fine particles or dust during transport and storage. This is an inconvenience
for the consumer and also a health hazard [12]. Fine particles and dust can also disturb feeding
systems of boilers and may lead to inhomogeneous combustion processes. Finally, dust may
contribute to fire and explosion risks during handling and storage [13].
The present paper aims to define an overview of the quality of wood pellet that Italian customers
could find on the market. Furthermore, the work aims to evaluate possible correlations of ash
content with other chemical parameters. This analysis has allowed to assess if the limits defined in
EN 14961-2 were chosen well and independently of each other.Analyses were carried out in the
Laboratorio Biomasse (Biomass Lab) of Università Politecnica delle Marche. Several pellet samples
were collected directly in the market avoiding any possible influence of producers and resellers on
the results. This kind of contribution and the related results are very difficult to find in literature.
2. Materials and methods
In this study, 88 bags of pellets were purchased directly from sale points distributed in Italy in
different regions during the period between November 2010 and February 2012. The choice of sale
points did not follow any criteria and neither the seller nor the manufacturer were informed about
the research. Within this investigation each pellet bag corresponded to one sample analyzed in the
Biomass Lab. For each sample analyses of some parameters considered by the EN 14961-2 were
carried out. The physical and chemical parameters taken into account in this investigation are
shown in table 1 and described as follows.
3
Table 1 – Chemical and physical parameters analyzed and corresponding reference methods
according to EN 14961-2 standard
Parameter Unit Normative references
Water content %1 EN 14774-2
Ash content %2 EN 14775
Mechanical durability %1 EN 15210-1
Diameter mm EN 14127
Length mm EN 14127
Net calorific value MJ kg-1 EN 14918
Bulk density kg m-3 EN 15103
Nitrogen %2 EN 15104
Sulphur mg kg-1 EN 15289
Chlorine mg kg-1 EN 15289
Potassium mg kg-1 EN 15290
Sodium mg kg-1 EN 15290
Manganese mg kg-1 EN 15290
Note 1: the percentage is a mass fraction of the sample as received
Note 2: the percentage is a mass fraction of dry matter
2.1 Determination of water content
The sample was dried at a temperature of 105 ± 2 °C in air atmosphere using forced ventilation
oven (mod. M120-VF, MPM Instruments) until constant mass is achieved and the percentage
moisture calculated from the loss in mass of the sample. The average water content was calculated
from two measurement series per sample.
2.2 Determination of ash content
The ash content is determined by calculation from the mass of the residue remaining after the
sample is heated in air under rigidly controlled conditions of time, sample weight and equipment
specifications to a controlled temperature of 550 ± 10 °C using a muffle furnace (mod. ZA, Prederi
Vittorio & figli). The average ash content was calculated from two measurement series per sample.
2.3 Determination of mechanical durability
The Mechanical durability tester, complying with the EN 15210-1, was used for the determination
of the mechanical durability. The test sample is subjected to controlled shocks by collision of pellets
against each other and against the walls of a specified rotating test chamber. In this tester
approximately 500 g of pellets are exposed to mechanical stress of 500 cycles of rotation of their
container. Before the test the pellets must be free of fines and therefore the share of fines must be
separated from the pellets by manual sieving using a 3.15 mm sieve. After the test the pellets are
weighed and the durability is calculated from the mass of sample remaining after separation of
abraded and fine broken particles. The average mechanical durability was calculated from two
measurement series per sample.
4
2.4 Determination of net calorific value
Isoperibolic calorimeter (mod.C2000 basic, IKA) was used to perform the sample combustion under
specific conditions in a bomb calorimeter according to EN 14918. The calorimeter was calibrated
with benzoic acid standard (IKA Benzoic Acid C723). The net calorific value is obtained by
calculation from the gross calorific value determined on the analysis sample. The calculation of the
net calorific value requires information about the water and hydrogen contents of the analysis
sample. These determinations were performed as well. The average net calorific value was
calculated from two measurement series per sample.
2.5 Determination of bulk density
The weight of a pellet sample has to be measured in a standard container of a given size and shape.
The volume of sample container (about 5.00 dm3) was determined by filling it with water. Bulk
density is calculated from the net weight per standard volume and reported for the measured
moisture content.
The average bulk density was calculated from three measurement series per sample.
2.6 Determination of nitrogen
A known mass of sample is burnt in oxygen/carrier gas mixture, under such conditions that it is
converted into ash and gaseous products of combustion. Oxides of nitrogen are reduced to nitrogen,
and those products of combustion which would interfere with the subsequent gas-analysis
procedures are removed. The carbon dioxide, water vapour and nitrogen mass fractions of the gas
stream are then determined quantitatively. The combustion, the following separation by a gas
chromatograph and detection of the elements were carried out by an Elemental Analyzer (mod.
2400 Series II CHNS/O System, Perkinelmer). Two determinations per sample were performed.
2.7 Determination of sulphur and chlorine
The sulphur and chlorine content was measured by decomposition in calorimetric bomb with excess
of oxygen and absorption of acid combustion gases in water (1 cm3). Detection of sulphate and
chloride was performed by liquid ion chromatography (mod. 761 COMPACT IC, Metrohm)
applying the principles of EN ISO 10304-1. The liquid ion chromatograph was calibrated using a
series of standards prepared from a Multielement Standards Solution 1,000 mg dm-3 (Fluka). Two
determinations per sample were performed.
2.8 Determination of potassium, sodium and manganese
The chemical analysis, carried out according to EN 15290, provides for the digestion of the biomass
sample in a closed container, using a mixture of acids (HNO3/H2O2/HF) and microwave oven (mod.
Multiwave 3000, Anton Paar). The mineralization product is then analyzed by inductively coupled
plasma optical emission spectrometry (ICP-OES mod. Optima 2100 PerkinElmer). The
spectrometer was calibrated using a series of standards prepared from a Single-Element Standard
1,000 mg dm-3 of Perkinelmer.
5
2.9 Determination of dimensions
The dimensions of pellets were determined according to EN 14127 by measuring the length and the
diameter of 20 randomly selected individual pellets per sample. The average values for the diameter
and the length were calculated.
2.10 Other data collected
From each pellet bag several information were also collected, in particular the producer/trader
country of origin and the eventual declaration of pellet origin (table 2). Furthermore for each bag a
sample was stored as backup.
Table 2 – Origin of the producer/trader and eventual declaration of pellet origin
Trader/producer location Declared origin
(n) (%) (n) (%)
Italy 45 51.1 43 48.9
Abroad 19 21.6 21 23.9
Unknown 24 27.3 24 27.3
TOTAL 88 100.0 88 100.0
2.11 Statistical analysis
A statistical analysis of results was carried out using Minitab Release 16 software for statistics. In
particular the analysis concerned:
- general descriptive statistic on all pellet samples;
- comparison of Net Calorific Values, chlorine, sulphur and nitrogen contents among pellet
samples classified on the base of ash content class level;
- multiple correlations among sulfur, nitrogen, chlorine, potassium, sodium, manganese and
ash contents on all pellet samples
- frequency analysis of ash, sulphur, chlorine and nitrogen with regards to EN 14961-2 A1
limit thresholds and analysis of the interdependencies of these parameters.
The subdivision of samples in groups of ash classes was based on range values established for the
A1, A2 and B classes as reported in EN 14961-2. In table 3 more details regarding the methodology
of this classification are shown.
Table 3 – Methodology used to classify the samples
Groups of samples Characteristics
Based on ash content level Samples group are divided on the base of ash content level
- A1 group: ash ≤ 0.7 %1
- A2 group: ash > 0.7 and ≤ 1.5 %1
- B group: ash > 0.7 and ≤ 3.0 %1
Note 1: the percentage is a mass fraction of dry matter
6
To investigate the relations between parameters, multiple correlations among sulfur, nitrogen,
chlorine, sodium, potassium, manganese and ash contents on all pellet samples were carried out
using Pearson correlation. The significance of calculated r values was assessed comparing the
calculated t value with the tabulated t value for 86 degrees of freedom. Further regression analysis
was carried out to point out the linkage between some element and ash content.
To assess the consistency of the A1 limit thresholds chosen in EN 14961-2 with the results of the
present work a series of frequency analyses were carried out. In particular we focused on the
hypothesis: "Once fuels reach or exceed the EN 14961-2 A1 limits for Cl and S, it is extremely
likely that also the ash content exceeds A1". To this end the frequency of conditions reported in
table 4 were evaluated and combined. To strengthen the outcome of the analysis some evaluations
were carried out considering even more strict limits for Cl (0,01) and S (0,01).
Table 4 – Conditions evaluated to assess the consistency of EN 14961-2 A1 limit thresholds for Cl
and S contents
Number Filter condition
1 Ash > 0.7
2 N > 0.3
3 Cl > 0.02
4 S > 0.03
5 N>0.3 or S>0.03 or Cl>0.02
6 Cl+S > 0.05
7 Cl > 0.01
8 S > 0.01
3. Results
The results of analyses carried out according to EN 14961-2 on all collected samples are reported in
table 4. This represents the general descriptive statistics on all pellet samples.
Table 5 – Results of analysis carried out on all the samples
Parameter Unit Mean
Standard
deviation Min Median Max
Water content %1 6.5 1.2 2.3 6.7 9.5
Net Calorific Value MJ kg-1 17.1 0.5 16.0 17.1 18.8
Ash content %2 0.7 0.6 0.1 0.5 2.5
Diameter mm 6.1 0.2 5.7 6.1 6.6
Length mm 17.2 2.4 11.9 17.1 24.4
Bulk density kg m-3 704.9 31.9 603.4 709.5 776.9
Durability %1 98.3 1.3 89.2 98.6 99.5
Nitrogen %2 0.19 0.27 0.02 0.11 1.99
Chlorine mg kg-1 39.0 77.9 0.5 19.9 531.4
Sulphur mg kg-1 79.7 52.2 23.2 62.2 269.7
7
Potassium mg kg-1 820.6 491.1 198.8 607.2 2184.6
Sodium mg kg-1 25.4 54.4 0.7 9.4 339.6
Manganese mg kg-1 76.5 36.2 0.4 69.3 209.7
Note 1: the percentage is a mass fraction of the sample as received
Note 2: the percentage is a mass fraction of dry matter
In general all mean values of the parameters analyzed in this investigation fall within the
requirements of EN 14961-2 for the A1 pellet quality class. Ash, nitrogen, chlorine, sodium and
sulphur contents show a higher variability than other parameters. The mean values are in general
similar to the median values as consequence of a regular distribution, apart from sodium and
chlorine. For all the elements, the mean is higher than the median. Anyway, despite considering the
variability and the range values for some parameters, several samples do not satisfy the
requirements established by A1 class. In table 5 are reported samples fulfilling the requirements of
the different quality classes according to EN 14961-2 with related percentages.
Table 6 – Samples that fulfill the requirements of the different quality classes according to EN
14961-2 with respect to all the parameters analyzed in this investigation
EN 14961-2 Class A1 A2 B Not compliant
Number of samples 45 17 15 11
Percentage 51.1% 19.3% 17.0% 12.5%
The subdivision of samples among ash classes groups entails that 45 samples are in A1 class, 17 in
A2 class and 15 in B class. Results of comparisons of Net Calorific Values, chlorine, sulphur and
nitrogen contents of pellet samples classified for ash content classes are reported in table 6.
Table 7 – Results of the comparisons of Net Calorific Value, chlorine, sulphur and nitrogen
content of pellet samples classified on the base of ash content class level
Parameter Unit Ash class Mean
Standard
deviation Min Median Max
Net Calorific Value A1 17.22 a 0.6 16.2 17.2 18.8
MJ kg-1 A2 16.96 a 0.4 16.2 17.1 17.7
B 16.88 a 0.5 16.0 17.1 17.6
Nitrogen A1 0.13 a 0.19 0.02 0.09 1.39
%1 A2 0.29 a 0.41 0.09 0.16 1.99
B 0.23 a 0.17 0.07 0.20 0.62
Chlorine A1 32.8 b 9.0 20.9 30.0 62.4
mg kg-1 A2 37.6 b 19.9 20.4 30.0 112.9
B 69.9 a 47.2 23.9 71.9 155.4
Sulphur A1 60.5 b 18.7 35.2 53.0 156.7
mg kg-1 A2 118.5 a 63.7 46.7 96.2 269.7
B 125.1 a 28.0 85.2 131.2 163.7
Note 1: the percentage is a mass fraction of dry matter
8
There is no significant difference among quality classes for Net Calorific Value and nitrogen
content. Regarding chlorine content there is a significant difference between B class and the other
two better classes. For sulphur content the significant difference is between the A1 class and the
other two worse classes.
Results of Pearson multiple correlations among sulfur, nitrogen, chlorine, sodium, potassium,
manganese and ash contents on all pellet samples are reported in table 7.
Table 8 – Correlation matrix with the results of the Pearson multiple correlation among nitrogen,
sulphur, chlorine, potassium, sodium, manganese and ash contents of all pellet samples
Ash Sulphur Chlorine Nitrogen Potassium Sodium Manganese
Ash 1
Sulphur 0.666*** 1
Chlorine 0.550*** 0.491*** 1
Nitrogen 0.306** 0.438*** 0.311** 1
Potassium 0.680*** 0.467*** NS 0.312** 1
Sodium 0.407*** 0.478*** 0.552*** 0.544*** NS 1
Manganese NS NS NS NS NS NS 1
***, **: Significant at p < 0.001 and p < 0.005 respectively
NS: Not Significant
Correlations among sulfur, nitrogen, chlorine and ash are all positive and significant. Potassium,
sulphur and chlorine have the higher values related to ash. Correlations of manganese with other
elements and ash content are all not significant. In general the positive correlations are moderate but
can give information on the most important elements. In particular, sulphur and potassium together
explain more than 60% of ash variance while all the analyzed parameters explain 80%.
The frequency analyses gave interesting results reported in table 8.
Table 9 - Results of the frequency analysis of conditions reported in table 4
Condition number Frequency
(n)
1 33
2 10
3 2
4 0
5 10
6 2
7 6
8 21
9
1 and 2 7
1 and 3 2
1 and 4 0
1 and 5 7
1 and 6 2
1 and 7 6
1 and 8 19
If ash exceeds A1 level, then 21%, 0% and 6% of these samples also exceed A1 levels repectively
for N, S and Cl. If N exceeds A1 level, then 70% of these samples also exceed A1 Ash levels S
never exceeds A1 level. If Cl exceeds A1 level , then all these samples (only 2) exceed A1 Ash
level. if N or S or Cl is above A1 level, then 70% of these samples also exceed A1 Ash level.
If the sum of Cl and S is above 0.05, then all these samples (only 2, the same as before) exceed A1
Ash level
The hypothesis: "Once fuels reach or exceed the A1 limits for Cl and S, it is extremely likely that
also the ash content exceeds A1" is confirmed but this happens with few samples because few
samples have high S (0) and Cl (2) content. To evaluate the significance of this finding, we
considered more strict limits for S and Cl and we still found that:
• If S exceeds A0 level , then all these samples (16 of 16) also exceed A1 Ash level
• If Cl exceeds A0 level , then all these samples (6 of 6) also exceed A1 Ash level
This means that the Cl and S limit thresholds of the standard are not consistent because too high
with respect to the ash limits. A0 level is still too high.
4. Discussion
The results of this work should be evaluated considering that Italian customers make use of wood
pellet mainly in small combustion devices for domestic heating like stoves and small boilers and
should prefer A1 class quality pellet. With respect to all the parameters analyzed in this
investigation emerges that only half of pellet samples fulfills all A1 class requirements. The quality
of analyzed pellet is thus far from being the best. The most critical parameter is the ash content and
all the samples fall within the limits scheduled in EN 14961-2 but only 51% within A1 group. The
statistical analysis of results pointed out moderate correlations between parameters. The most
important elements linked to the ash content are potassium and sulphur. Other correlations with ash
and between elements are significant but less intense. Nitrogen in particular has mainly an organic
origin and is only partially linked to other elements and ash content.
Referring to the A1 class of virgin biomass pellets, the results of the frequency analysis would seem
to indicate pellets that samples within the ash limit are well below the limits of chlorine and sulfur.
10
However, the need to identify treated wood and the difficulties of chlorine and sulfur detection at
very low contents, do not allow us to express an opinion on the possible reduction of the current
thresholds.
5. Conclusions
The quality of pellet samples considered in this investigation is not fully satisfactory if compared
with the characteristics defined by the European standard. Only half of the samples collected from
the market meets the quality requirements of EN 14961-2 and more than 10% of these are outside
of the standard. Differences are important above all for ash content, sulphur content and durability,
that suggests a greater attention in manufacturing process and selection of raw material. The study
also shows that samples belonging to group A1 and A2 have lower sulfur and chlorine contents.
The results of the correlation analysis show how ash is a very important quality parameter to be
monitored because it is linked in particular to elements like sulphur, chlorine and potassium,
representing problems for combustion devices, especially in terms of corrosion. Relativamente ai
parametri analizzati in questo lavoro, l’ulteriore miglioramento qualitativo del pellet dovrebbe
essere legato più sul livello di ceneri che non sugli altri parametri.
With regards to the parameters analyzed in this work, the further improvement of pellet quality
should be tied more on the ash level than on the other parameters.
References
[1] Sikkema S, Steiner M, Junginger M, Hiegl W, Hansen MT, Faaj A. The European wood
pellet markets: current status and prospects for 2020. Biofuel Bioprod Bior 2011;5(3):28.
[2] Selkimäki M, Mola-Yudego B, Röser D, Prinz R, Sikanen L. Present and future trends in
pellet markets, raw materials, and supply logistics in Sweden and Finland. Renew Sust Energ Rev
2010;14(9):3068-75.
[3] Biswas AK, Yang W, Blasiak W. Steam pretreatment of Salix to upgrade biomass fuel for
wood pellet production. Fuel Process Technol 2011;92(9):1711-7.
[4] Monteiro E, Mantha V, Rouboa A. Portuguese pellets market: Analysis of the production
and utilization constrains. Energ Policy 2012;42(0):129-35.
[5] Stahl M, Berghel J. Energy efficient pilot-scale production of wood fuel pellets made from a
raw material mix including sawdust and rapeseed cake. Biomass Bioenerg 2011;35(12):4849-54.
[6] Filbakk T, Jirjis R, Nurmi J, Høibø O. The effect of bark content on quality parameters of
Scots pine (Pinus sylvestris L.) pellets. Biomass Bioenerg 2011;35(8):3342-9.
[7] Olsson O, Hillring B, Vinterbäck J. European wood pellet market integration – A study of
the residential sector. Biomass Bioenerg 2011;35(1):153-60.
[8] EN 14961-1. Solid biofuels - Fuel specifications and classes - Part 1: General requirements.
Brussels: CEN; 2010.
[9] EN 14961-2. Solid biofuels - Fuel specifications and classes - Part 2: Wood pellets for non-
industrial use. Brussels: CEN; 2011.
[10] Werkelin J, Skrifvars B-J, Zevenhoven M, Holmbom B, Hupa M. Chemical forms of ash-
forming elements in woody biomass fuels. Fuel 2010;89(2):481-93.
11
[11] Ghafghazi S, Sowlati T, Sokhansanj S, Bi X, Melin S. Particulate matter emissions from
combustion of wood in district heating applications. Renew Sust Energ Rev 2011;15(6):3019-28.
[12] Vinterbäck J. Pellets 2002: the first world conference on pellets. Biomass Bioenerg
2004;27(6):513-20.
[13] Lehtikangas P. Storage effects on pelletised sawdust, logging residues and bark. Biomass
Bioenerg 2000;19(5):287-93.
