Author’s Accepted Manuscript

Determination of silicon in biomass and products of

pyrolysis process via high-resolution continuum
source atomic absorption spectrometry

Flávio V. Nakadi, Caroline Prodanov, Wiliam

Boschetti, Maria Goreti R. Vale, Bernhard Welz,
Jailson B. de Andrade
www.elsevier.com/locate/talanta

PII: S0039-9140(17)31224-9
DOI: https://doi.org/10.1016/j.talanta.2017.12.022
Reference: TAL18158
To appear in: Talanta
Received date: 12 September 2017
Revised date: 6 December 2017
Accepted date: 7 December 2017
Cite this article as: Flávio V. Nakadi, Caroline Prodanov, Wiliam Boschetti,
Maria Goreti R. Vale, Bernhard Welz and Jailson B. de Andrade, Determination
of silicon in biomass and products of pyrolysis process via high-resolution
continuum source atomic absorption spectrometry, Talanta,
https://doi.org/10.1016/j.talanta.2017.12.022
This is a PDF file of an unedited manuscript that has been accepted for
publication. As a service to our customers we are providing this early version of
the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting galley proof before it is published in its final citable form.
Please note that during the production process errors may be discovered which
could affect the content, and all legal disclaimers that apply to the journal pertain.
 Determination of silicon in biomass and products of pyrolysis process via high-
resolution continuum source atomic absorption spectrometry

Flávio V. Nakadia*, Caroline Prodanova, Wiliam Boschettia, Maria Goreti R. Valea,b**,

Bernhard Welzb,c, Jailson B. de Andradeb
a
Instituto de Química, Universidade Federal do Rio Grande do Sul, Porto Alegre, RS,
Brazil
b
Instituto Nacional de Ciência e Tecnologia do CNPq–INCT de Energia e Ambiente,
Universidade Federal da Bahia, Salvador, BA, Brazil
c
Departamento de Química, Universidade Federal de Santa Catarina, Florianópolis, SC,
Brazil
flavio.nakadi@ufrgs.br
mgrvale@ufrgs.br
*Corresponding authors:

1
ABSTRACT

Thermochemical processes can convert the biomass into fuels, such as bio-oil. The
biomass submitted to pyrolysis process, such as fibers, are generally rich in silicon, an
element that can lead to damages in an engine when there is high concentration in a
fuel. High-resolution continuum source atomic absorption spectrometry (HR-CS AAS)
is an interesting alternative for Si determination in the products and byproducts of the
pyrolysis process because, besides the flame (F) and graphite furnace (GF) atomizers, it
has enhanced the application of direct analysis of solid samples (SS) within GF. This
study aimed the development of methods to determine Si in biomass samples, their
products and byproducts using HR-CS AAS. A high-resolution continuum source
atomic absorption spectrometer contrAA 700 equipped with F and GF atomizers was
used throughout the study. HR-CS F AAS ( = 251.611 nm, 1 detection pixel,
N2O/C2H2 flame) was used to evaluate Si content in biomass and ash, after a
microwave-assisted acid digestion with HNO3 and HF. HR-CS GF AAS (Tpyr = 1400
°C, Tatom = 2650 °C) has evaluated Si in pyrolysis water and bio-oil at 251.611 nm, and
in peach pit biomass and ash at 221.174 nm using SS, both wavelengths with 1
detection pixel. Rhodium (300 μg) was applied as permanent modifier and 10 μg Pd + 6
μg Mg were pipetted onto the standards/samples at each analysis. Three different
biomass samples were studied: palm tree fiber, coconut fiber and peach pit, and three
certified reference materials (CRM) were used to verify the accuracy of the methods.
The figures of merit were LOD 0.09 - 20 mg kg–1, and LOQ 0.3 - 20 mg kg–1,
considering all the methods. There were no significant differences between the CRM
certified values and the determined ones, using a Student t-test with a confidence
interval of 95% (n = 5). Si concentration ranged from 0.11 to 0.92% m m–1, 1.1 to 1.7
mg kg–1, 3.3 to 13 mg kg–1, and 0.41 to 1.4% m m–1, in biomass, bio-oil, pyrolysis water
and ash, respectively. Si remained mostly in the ash, leading to a mass fraction of up to
103%, even when the Si loss is not considered. Silicon concentration in bio-oil was
below 1.7 mg kg–1, which is suitable for its application as a fuel. The developed
methods using HR-CS AAS are suitable for Si determination in biomass, bio-oil,
pyrolysis water, and ash. The application of bio-oil as an alternative fuel would be
possible evaluating its Si content due to its low levels. The mass balance for Si has
proved to be an important tool in order to evaluate the correct disposal of pyrolysis
process byproducts.
2
Keywords: Biomass; Bio-oil; Silicon determination; High-resolution continuum source
atomic absorption spectrometry; Mass balance.

1. Introduction

The use of alternative sources for energy production has increased significantly
in the last decades. Alternative energy encompasses sunlight, wind, tides, rain,
geothermal heat, biomass, etc., which are widely available and environment friendly as
they cause little or almost no pollution [1-3]. The Biomass Magazine, for example,
reported that in 2015 almost 25% of the electricity generated in the UK came from
renewable sources representing a growth of 29% when compared with 2014 [2].
According to the International Energy Agency, 15% of the US electricity generated in
2016 was obtained from renewable sources, whereas in Brazil, in the same year, almost
84% of the electricity was produced from renewable sources [3].
Biomass represents approximately 10% of the renewable energy produced in
Brazil [3]. It is defined as all the organic matter obtained from plants, even that usually
treated as waste material from other processes [4]. Biomass can be converted into
energy through different thermochemical processes. The type and quantity of biomass
available, the desired form of energy, the environmental regulations and the economic
conditions are the decisive factors used to define which thermochemical process should
be employed. These processes can be classified into two categories: (i) the direct
combustion process, involving the direct use of biomass as fuel and (ii) processes that
convert biomass into other useful energetic products before their use as an energy
source. This latter category includes the gasification and pyrolysis processes [5,6].
Pyrolysis offers a flexible and attractive way of converting solid biomass into a
liquid that can be easily stored and transported, and which can be successfully used to
produce heat, power, and/or chemicals. The pyrolysis process is a thermal
decomposition operated in the absence of oxygen, in which biomass is rapidly heated
and converted mostly into condensable vapors and aerosols which, after cooling and
condensation, give rise to a liquid product. Generally, this liquid product comprises two
phases, a non-aqueous phase, called bio-oil, consisting of high-molecular-weight
molecules, and an aqueous phase, which consists of low-molecular-weight compounds.
This aqueous phase comes from the moisture present in the biomass and from products
3
of reactions that occur during the pyrolysis process. The pyrolysis process (and the
direct combustion process) produces also a solid byproduct, the ash, which has to be
considered [5-9].
As a renewable liquid fuel, bio-oil can serve as a substitute for fuel oil and diesel
in many static engines, including boilers, furnaces, and turbines for electricity
generation. Alternatively, the bio-oil can be used as a raw material for the production of
lubricants, adhesives, phenol-formaldehyde-type resins, wood flavors, etc. [10]. On the
other hand, the pyrolysis aqueous phase and the ash have no application; in the best
case, they can be used as fertilizer in agriculture, depending on their composition.
It is known that the presence of high amounts of Si in fuels can lead to damages
in the internal combustion chamber of an engine [11,12]. Nevertheless, in fertilizers Si
should be in high concentration as it is considered an essential element for the plant
development and health [13,14]. There is no legislation for the presence of Si in bio-oil
samples. The Brazilian National Agency of Petroleum, Natural Gas and Biofuels,
published a resolution [15], which establishes a maximum concentration of 80 mg kg–1
Si plus Al in diesel. In the case of the ash, the Brazilian legislation [16] reports a
minimum concentration of 1% m m–1 of Si in solid fertilizers.
In this context, the mass balance can be considered an important tool to evaluate
the distribution of the elements present in the biomass samples through the obtained
products and byproducts. Some studies have determined the mass balance of lead [17],
chromium and cadmium [18] in biomass, bio-oil, pyrolysis water and ash. The objective
was to verify the behavior of these elements during the pyrolysis process thus
evaluating the bio-oil quality (in terms of the analytes) and the proper discard or use of
the byproducts. Therefore, fast and reliable analytical methods are required in order to
perform the determination of Si in biomass, bio-oil, and byproducts.
The most commonly used analytical techniques for the analysis of these matrices
are the inductively coupled plasma optical emission spectrometry (ICP OES) [19-21],
the line source atomic absorption spectrometry (LS AAS) [22,23] and the high-
resolution continuum source atomic absorption spectrometry (HR-CS AAS) [24-27].
There is nothing in the literature about the determination of Si in bio-oil. Considering
similar materials, Oliveira et al. [27] determined Si in vegetable oil and biodiesel by
HR-CS AAS using a flame as the atomizer (HR-CS F AAS) and a sample dilution with
xylene; and Amais et al. [28] compared a dilute-and-shoot method to analyze diesel and
biodiesel with microwave-assisted acid digestion and microemulsion using microwave-

4
induced plasma OES for Si determination. Luz and Oliveira [29] determined Si in
petrochemical products (crude oil, diesel or gasoline) using an emulsion with Triton X-
100 and hexane (only for crude oil); the emulsions were analyzed using LS AAS with a
graphite furnace as atomizer (LS GF AAS). Gazulla et al. [30] determined Si in naphtha
samples using ICP OES and isooctane as the diluent. It is noteworthy the cited studies
usually required sample pretreatment procedures.
Several studies can be found in the literature reporting the determination of Si in
biomass [20,21,26]. Le Blond et al. [20] developed a method for the determination of Si
in biomass samples by ICP OES after a microwave-assisted acid digestion with HF.
Tafur-Marinos et al. [21] investigated different digestion procedures for the analysis of
wood (as biomass) by ICP OES. Seventeen elements were evaluated, including Si, after
wet and dry fusion of the ash, and microwave-assisted digestion of the wood samples.
The ash fusion was the most accurate for the determination of Si. Boschetti et al. [26]
developed a method for the determination of Si in plant material by HR-CS GF AAS
and direct analysis of solid samples (SS).
HR-CS AAS appears to be a good alternative for the determination of Si in
biomass, its products and byproducts. With this technique, it is possible to analyze
samples with high Si contents, using the HR-CS F AAS, and samples with Si at trace
levels, using the HR-CS GF AAS. Due to the use of a high-intensity xenon short-arc
lamp as radiation source, it is possible to perform measurements at any wavelength
between 190 and 900 nm. The combination of this lamp with the high-resolution double
monochromator and the linear charge-coupled device (CCD) detector [31] provides
better limits of detection (LOD) and an extended linear working range [32,33]. HR-CS
GF AAS also makes possible the direct analysis of solid and liquid samples [34].
Therefore, only a minimum of pretreatment procedures is required for sample
preparation. This is an advantage because less time and reagents are required,
decreasing the cost and the possibility of sample contamination. Furthermore, when
trace analysis is required, the sensitivity is improved through the HR-CS SS-GF AAS
since dilutions are avoided [35,36]. The HR-CS F AAS has an accessory, a flow
injection system SFS6, which allows greater reproducibility, and lower sample
consumption. Moreover, the sensitivity can be adjusted controlling the sample volume
injected in each analysis [37-40].
The present work shows the development of methods to determine Si in biomass
samples, their products and byproducts using HR-CS AAS. The use of flame and

5
graphite furnace atomizers was evaluated according to the Si content in the samples.
The concentration and mass balance of Si were investigated in the biomass, bio-oil,
pyrolysis aqueous phase and ash.

2. Experimental

2.1. Instrumentation

All the measurements were carried out with a high-resolution continuum source
atomic absorption spectrometer contrAA 700 (Analytik Jena AG, Jena, Germany)
equipped with a flame and a transversely heated graphite tube atomizer. The optical
system consists of a 300 W xenon short-arc lamp (GLE, Berlin, Germany) as the
radiation source operating in a hot-spot mode, a high-resolution double monochromator
which comprises a Littrow prism monochromator followed by an echelle grating
monochromator, and a linear CCD array detector with 588 pixels, 200 of which are used
for analytical purposes.
The analytical lines at 251.611 nm (100% relative sensitivity) and 221.174 nm
(5.9% relative sensitivity) were evaluated with 1 pixel detection (center pixel only) for
Si determination. An N2O/acetylene flame (2:1 flow ratio) and a flow-injection valve
SFS6 (Analytik Jena AG) were used in HR-CS F AAS to evaluate biomass and ash
samples. The injection valve enables a controlled injection of small sample volumes,
resulting in an absorbance peak, which can be evaluated by its area. The sample flux
was 6 mL min–1; 0.5 s sample injection and 10 s integration time were used with this
device. HR-CS GF AAS was used for Si determination in pyrolysis water and bio-oil,
and the heating program is shown in Table 1. All determinations have been carried out
using pyrolytically coated SS graphite tubes and SS graphite platforms (Analytik Jena
AG). Argon 99.996% (White Martins, São Paulo, Brazil) was used as purge gas through
all temperature steps, except during atomization when the Ar flow was interrupted. The
samples were weighed onto the graphite platforms using an M2P microbalance
(Sartorius, Göttingen, Germany), and they were transferred manually into the graphite
tube with a pre-adjusted pair of tweezers, which are part of an SSA 6 manual solid
sampling accessory (Analytik Jena AG).

Table 1

6
2.2. Reagents and solutions

All reagents were of analytical grade or higher purity. Deionized water obtained
from a Milli-Q water purification system (Millipore, Bedford, USA) with a specific
resistivity of 18.2 MΩ cm was used to prepare solutions and standards, dilution and
cleaning. The standard solutions were prepared by diluting a 1000 mg L–1 Si stock
standard solution (Merck, Darmstadt, Germany) in polypropylene ware and flasks. They
were cleaned by soaking them in a 1.4 mol L–1 HNO3 solution overnight, further rinsing
with deionized water, and finally dried in a laminar flow hood.
Rhodium was used as permanent chemical modifier by ten serial injections of 30
µL each of a 1000 mg L–1 Rh stock solution (Alfa Aesar, Parkridge, USA). Each
injection was followed by a heating program shown in Table 2, resulting in a deposition
of a final mass of 300 μg Rh on the platform surface [26]. In addition to this permanent
chemical modifier, 10 μL of a chemical modifier in solution, consisting of 0.1% m v–1
Pd, 0.06% m v–1 Mg, and 0.05% m v–1 Triton X-100 (solutions from Sigma-Aldrich)
were injected together with the sample and standard solutions.

Nitric acid (Anidrol, São Paulo, Brazil) was previously purified by sub-boiling
distillation in a quartz still (Marconi, Piracicaba, Brazil) and used for cleaning and for
acid digestion; hydrofluoric acid (Merck) and boric acid (Merck) were also used for
acid digestion. Bio-oil samples were diluted in propan-1-ol (Merck) before analysis.

2.3. Samples and certified reference materials

Three different biomass samples were collected for this study: palm tree fiber,
coconut fiber and peach pit, all of them acquired as waste from industries located in the
northeast and south of Brazil.
Bio-oil was extracted from them by a fast pyrolysis process, in which the
samples were heated to 700 °C under N2 atmosphere in a lab-made vertical oven,
described by Moraes et al. [41] for 10 min. The bio-oil is collected together with the
pyrolysis water, and the ash remains inside the reactor. There was no system to collect

7
particulates or gases generated in this process. The ash was collected directly from the
reactor; a liquid-liquid extraction with chloroform separated the bio-oil from the
pyrolysis water. All fractions were weighed for mass balance evaluation, including the
biomass before the pyrolysis process.
Three certified reference materials (CRM), NCS DC 73349 Bush branches and
leaves (China National Analysis Center for Iron & Steel, Beijing, China) for biomass,
SRM 1084a Wear-metals in lubricating oil (NIST, Gaithersburg, USA) for bio-oil, and
BCSS-1 Marine sediment (NRC-CNRC, Ottawa, Canada) for ash, were used to verify
the accuracy of the methods.

2.4. Sample preparation

Biomass and ash samples were ground in a micro-mill A-11 Basic (IKA-Werke,
Staufen, Germany) and dried at 65 ± 5 °C in a stove for at least 3 h. Later, the samples
were sieved through a 200 μm polyester sieve and stored in plastic vials to avoid silicon
contamination. Bio-oil and pyrolysis water were also stored in plastic vials, and the bio-
oil was heated to 40 °C for 40 min, stirred for 30 s to homogenize its content, and
cooled before its analysis. No preparation was needed for pyrolysis water.
Depending on their Si content, biomass and ash samples were either analyzed by
HR-CS SS-GF AAS with no further preparation, or by HR-CS F AAS. An acid
digestion method was developed for the F AAS determination. The optimized
conditions of the microwave-assisted oven Multiwave PRO (Anton Paar, Graz, Austria)
are shown in Table 3; for a sample mass of 100 mg, 5.0 mL HNO3, and 1.5 mL HF were
used, and after the digestion 3.0 mL of a saturated H3BO3 solution was added.

The bio-oil samples had to be diluted before analysis using HR-CS GF AAS due
to their Si concentration. Propan-1-ol was selected as the solvent (the bio-oil and
propan-1-ol were mixed by shaking them manually until only one phase was observed),
and its volume varied depending on the Si content in the samples. The dilution factor
ranged from 1:4 to 1:10 m m1 (bio-oil:propan-1-ol). It is noteworthy the dilution was
easier to handle comparing with raw bio-oil. Both diluted bio-oil (weighing 4–10 mg)
and pyrolysis water (10 L) were pipetted manually directly onto the graphite platform.
8
3. Results and discussion

3.1. Sample preparation

During the pyrolysis process, Si from the biomass is distributed through the
three collected fractions: bio-oil, pyrolysis water and ash. It was found that Si
concentration was higher in biomass and ash, and lower in bio-oil and pyrolysis water.
Due to this difference, both F AAS and GF AAS and different sample preparation
techniques were needed to achieve a reliable determination of silicon. The optimization
included all the three biomass samples (coconut fiber, palm tree fiber, and peach pit)
and their pyrolysis products and byproducts.
The most common approach of sample preparation for F AAS is an acid
digestion. Silicon is known as a tricky element to be released from the matrix due to its
tendency to form silicates, which are insoluble in water. The use of hydrofluoric acid is
a common practice in this case, as Si has a higher affinity to fluorine than to oxygen,
which results in the formation of silicon fluoride complexes, such as SiF62–. On the
other hand, the presence of HF could also result in the formation of volatile species,
such as SiF4, that are easily lost in the digestion process, which requires special care.
The microwave-assisted acid digestion has some advantages for this purpose: (i)
the heating is controlled by a temperature sensor, as well as the temperature ramp; (ii)
the system is closed, which would avoid losses of volatile species; and (iii) the PTFE
vessels enable the use of HF. A four-step heating program (Table 3) was chosen to
minimize the formation of volatile species. The temperatures and hold time, as well as
the volumes of HNO3 and HF, were optimized for the biomass and ash samples. For
instance, 0.5, 1.0, 1.5, and 2.0 mL of HF were evaluated with CRM BCSS-1. After each
digestion, the resulting liquid was analyzed by HR-CS F AAS until the absorbance
signal reached a maximum and remained constant: there was no difference of Si
concentration comparing 1.5 and 2.0 mL of HF, the former used throughout the study.
The result of this optimization has been described in Section 2.4.
Direct solid sample analysis has been selected for HR-CS GF AAS, thus
minimum sample preparation was necessary. The content of Si in bio-oil was elevated,
and its viscosity restricted pipetting onto the graphite platform. Therefore, the bio-oil
samples were diluted with propan-1-ol according to their Si content.

9
3.2. Optimization of the HR-CS F AAS conditions

The ratio of N2O/C2H2 and reading height were optimized automatically by the
equipment software in order to obtain the highest absorbance value. The conditions
were: 190 L h–1 of acetylene, 370 L h–1 of nitrous oxide, 8 mm burner height with a 50
mm burner slot. The sample flux (inlet) was optimized manually, and 6 mL min–1 was
chosen using the same parameter. The measurements were carried out at 251.611 nm.
The equipment has a flow-injection valve, which results in a peak-shaped signal,
enabling the evaluation of peak area instead of the height (commonly used in F AAS),
hence improving the repeatability between readings. Moreover, flow-injection systems
allow smaller sample volume intake, a great advantage in a technique that usually
requires 2–3 mL of sample. This valve changes its intake from cleaning water to
sample, lowering variations in the flame temperature. Therefore, two parameters are
important to evaluate: injection time and load time.
Injection time corresponds to the sample injection into the nebulizer (there is no
sample loop in this system), and load time is the time needed to fill the system with
sample. During measurement, the sample is injected for a selected period of time, and
after that the valve switches its inlet to pass a carrier (deionized water). As the Si
concentration in the samples was high, 0.5 s of injection time was sufficient, with a load
time of 5 s. The peak profile is shown in Fig. 1, and the integration time was set to 10 s
to observe the entire peak properly.

3.3. Graphite furnace temperature program

The biomass and ash of the peach pit had a Si concentration below the limit of
detection for HR-CS F AAS. Therefore, a method for HR-CS SS-GF AAS has been
developed for this determination.
First, the method was evaluated for pyrolysis water and bio-oil analysis, as these
matrices are simpler than solid samples. A method developed previously by our
research group [26] has been employed for Si determination. Two differences were
needed: increasing the time of the drying step to ensure complete removal of the

10
solvent, and no Ar flow was needed during atomization. Two chemical modifiers were
employed: (i) 10 μL of a Pd/Mg/Triton X-100 solution (0.1% m v–1, 0.06% m v–1, and
0.05% m v–1, respectively) was injected over the standards/samples in each analysis;
and (ii) 300 μg Rh as permanent modifier, which was added on the SS graphite platform
surface before the analysis. The temperature program for coating the SS platform with
Rh can be found in Table 2. The coating was stable for, approximately, 150
measurements.
Although the method described by Boschetti et al. [26] was developed for
biological samples, the optimization of pyrolysis and atomization temperatures for
peach pit biomass and ash was carried out in this study at 221.174 nm, because the
concentration of Si has been considerably higher in these matrices than in bio-oil and
pyrolysis water. The pyrolysis and atomization curves are shown in Fig. 2.

The method should be robust for all the analyzed matrices, including the Si
standard solution, in order to find a compromise condition. Based on Fig.2, the
pyrolysis temperature of 1400 °C has provided the best results considering the three
matrices (standard, biomass, and ash). Although the pyrolysis temperature has led to
higher integrated absorbance (Aint) at lower temperatures (< 600 °C) for biomass, it was
due to particulate matter generated from the matrix, which scattered the radiation. The
pyrolysis curve for ashes has no significant difference between 1000 and 1400 °C, the
highest Aint values observed in the optimization. Both matrices have generated solid
residues after analysis, and this effect was minimized applying higher pyrolysis
temperatures. Pyrolysis temperatures higher than 1600 °C caused a decrease in Aint
values.
The atomization temperature of 2650 °C has the highest Aint for the Si standard
solution. There were no significant differences in Aint ranging from 2400 to 2650 °C for
biomass and ash. However, 2650 °C has presented lower relative standard deviation
(RSD). Therefore, 2650 °C was chosen as the atomization temperature, which is in
agreement with the literature [26]. The temperature program used for Si determination
in biomass, ash, pyrolysis water and bio-oil is described in Table 1. No spectral
interference was observed throughout the study (Fig. 3).

11
3.4. Influence of sample mass in HR-CS SS-GF AAS

The investigation of the sample mass that is possible to analyze, maintaining a

linear relation with the integrated absorbance of Si, is important to evaluate. Both,
biomass and ash samples, had good linearity (Rbiomass = 0.9894 and Rash = 0.9928) for a
mass range from 0.05 mg to 0.35 mg. Therefore, this mass range has been used
throughout the analysis (Fig. 4).

3.5. Figures of merit

The figures of merit of the methods are summarized in Table 4. Three

calibration curves were used with aqueous Si standard solutions: (i) at 251.611 nm for
HR-CS F AAS, from 10–80 mg L–1 Si (R = 0.9981) and (ii) for HR-CS GF AAS from
0.5–5.0 ng Si (R = 0.9973); (iii) at 221.174 nm for HR-CS SS-GF AAS from 20–100 ng
Si (R = 0.9993). The LOD was calculated using three times the standard deviation of ten
measurements of a blank solution, divided by the slope of the calibration curve
(sensitivity), and the limit of quantification (LOQ) ten times the same standard
deviation, divided by the sensitivity. Blank measurements were carried out according to
the “zero mass response” principle, which states the blank in SS could be evaluated as
the absence of the sample matter. It is noteworthy the blank measurements had the
chemical modifiers onto the graphite platform, as well as propan-1-ol for the bio-oil
analysis, as described previously. The concentration in mg kg–1 was obtained
considering 100 mg of sample digested in a final volume of 15 mL for HR-CS F AAS;
using HR-CS GF AAS, for bio-oil and water samples at 251.611 nm, the highest mass
weighed onto the graphite platform was 10 mg, and for the solid samples at 221.174 nm
it was 0.1 mg. Characteristic concentration (c0) and mass (m0) are defined as the
concentration/mass of the analyte that corresponds to an integrated absorbance (in this
case, for both atomizers) of 0.0044 s.

12
 The figures of merit were first compared with the ones obtained by Boschetti et
al. [26] for the analysis of plant material using HR-CS SS-GF AAS at 221.174 nm,
which are similar to the method developed in this work for biomass and ash (peach pit
sample), with a LOD of 5.0 mg kg–1, a LOQ of 17 mg kg–1, and a characteristic mass of
2.0 ng. Kowalewska et al. [25] evaluated the behavior of Si in aqueous and organic
media using HR-CS F AAS, and they have reported an LOD of 0.05 mg L–1 or < 0.01
mg L–1, depending on the Si compound, in organic solutions. In aqueous phase, the best
LOD was 0.23 mg L–1, slightly higher than the one found by Gómez-Nieto et al. [40]
(0.16 mg L–1). Comparing the method for bio-oil using HR-CS GF AAS, the figures of
merit were comparable with a method using line source AAS for petrochemical
products [29], with a LOD of 0.08 mg kg–1, and a LOQ of 0.26 mg kg–1 at 251.6 nm,
using niobium carbide as permanent modifier.

3.6. Evaluation of the accuracy

Three CRM were used to evaluate the accuracy of the methods: NCS DC 73349
for biomass, SRM 1084a for bio-oil, and BCSS-1 for ash. Pyrolysis water was
compared with an aqueous standard solution directly. They were analyzed using all the
methods described previously: acid digestion for biomass and ash using HR-CS F AAS;
bio-oil dilution in propan-1-ol and SS for peach pit biomass and ash using HR-CS GF
AAS; the results are presented in Table 5. There were no significant differences
between the certified values and the determined ones, using a Student t-test with a
confidence interval of 95% (n = 5).

It would be possible to determine Si using HR-CS GF AAS in all the samples;

however, the RSDs were around 10%, i.e., higher when compared to HR-CS F AAS.
Therefore, HR-CS F AAS is considered to be more suitable for the analysis of biomass
and ash.

3.7. Sample analysis and mass balance

13
 The methods are summarized in Table 6. Three samples were studied: coconut
fiber, palm tree fiber, and peach pit, and their pyrolysis products and byproducts which
were obtained by the process described in Section 2.3. The samples have to be weighed
before the process in order to obtain the total mass of Si, and further bio-oil, pyrolysis
water, and ash are also weighed to understand the distribution of the analyte (silicon)
during the process. Two previous papers have demonstrated this possibility for Pb [17],
Cd and Cr [18], and it is possible to calculate the Si retention in each fraction (related to
the total mass of Si, originally present in the biomass) by using equation 1, where
Cfraction and Cbiomass are the Si concentration in the evaluated fraction (bio-oil, water, or
ash) and in the biomass, respectively, with the same units, and mfraction and mbiomass are
the mass (in g) of fraction obtained in the pyrolysis process and the mass (in g) of
biomass used in the pyrolysis process, respectively. The Si concentration in all pyrolysis
fractions for the three samples and the mass balance are shown in Table 7.

mfraction
( fraction )
m iomass
Retention in raction / 100 (1)
iomass

Several difficulties might occur when acquiring the different fractions: non-
efficient separation of bio-oil and water, losses due to volatilization, etc. The sum of the
Si content in all fractions might not always be 100%, and this difference is called “loss”
in Table 7. Evaluating the palm tree fiber sample, 57% of Si was not recovered,
probably due to some kind of volatilization. Coconut fiber and peach pit, on the other
hand, had good recoveries, i.e., lower than 7% in the pyrolysis process.
Low silicon content is desirable in fuels, and the bio-oils investigated in this
work fit this requirement. Their concentration was below 2 mg kg–1 in all samples,
which corresponds to a mass fraction of less than 0.02%. This is an important
information, considering the high concentration of Si in the biomass samples (between
0.1 and 0.9% m m–1). Pyrolysis water also had low Si content, below 15 mg kg–1, which
is interesting for its disposal. As expected, Si remained mostly in the ash, probably as

14
refractory silicates, leading to a mass fraction of up to 99%, even when the Si loss in
peach pit is not considered. The concentration of Si in the three ash samples, between
0.4 and 1.4 % m m–1, would be useful for industries such as ceramics and construction.
Moreover, the pyrolysis water and ash together could be interesting for fertilization as
slurry, combining organic matter and mineral sources, respectively.
The tendency observed in Table 7 is straight, evaluating the Si content in the
fractions of all three samples. It might be assumed that other biomass samples would
behave the same way during a pyrolysis process: Si would remain in ash, leaving the
bio-oil almost free of this element, which is crucial for its potential application as an
alternative fuel.
Clearly, Si shows a different behavior when compared to Cd, Cr, and Pb [17,18].
There were severe losses of these elements during the pyrolysis process, ranging from
62% to 100%, which could be expected for Cd and Pb due to their volatility. Cr, in
contrast, is considered a refractory element, which could be expected to be found mostly
in the ash. Depending on the matrix composition, some volatile species of Cr, such as
CrCl6 and CrOCl4, could be formed at 700 °C and Cr also tends to associate with
particulate matter, behaviors that would not affect significantly Si [42]. Nevertheless,
we have to realize that there was a 57% loss of Si from the palm tree fiber biomass,
which means that there might be loss mechanisms that are not yet fully understood.

4. Conclusions

The developed methods using HR-CS AAS with both, flame and graphite
furnace atomization, including direct solid sample analysis, are suitable for Si
determination in biomass, bio-oil, pyrolysis water, and ash. Although HR-CS F AAS
required a sample preparation (microwave-assisted acid digestion in this study), it
proved to be reliable, fast, and accurate for Si determination in biomass and ash. HR-CS
GF AAS was versatile to achieve the goal of this study in all investigated matrices,
distinct from each other, which corroborates the great potential of this technique.
Furthermore, the application of bio-oil as an alternative fuel would be possible
evaluating its Si content due to its low levels. The knowledge of Si concentration in
both pyrolysis water and ash is significant, because proper decisions could be made
considering their final destinations. The mass balance for Si has proved to be an

15
important tool in a pyrolysis process, which could predict the composition of the
fractions of a biomass pyrolysis.

Acknowledgements

The authors are grateful to Conselho Nacional de Desenvolvimento Científico e

Tecnológico (CNPq) for financial support. F.V.N, C.P, W.B., M.G.R.V., B.W. and
J.B.A. have research scholarships from CNPq (Process nos. 150250/2016-9,
144108/2016-0, 121824/2017-9, 305679/2015-5, 303526/2016-5 and 305619/2014-4,
respectively).

References
[1] M. Hasanuzzaman, U.S. Zubir, N.I. Ilham, H.S. Che, Global electricity demand,
generation, grid system, and renewable energy polices: a review, WIREs Energy
Environ. 6 (2017) 1-18.
[2] Biomass Magazine, Renewable energy statistics. www.biomassmagazine.com, 2017
(accessed 20.07.17).
[3] International Energy Agency, Renewable energy statistics. www.iea.org, 2017
(accessed 20.07.17).
[4] P.K. Kanaujia, Y.K. Sharma, M.O. Garg, D. Tripathi, R. Singh, Review of analytical
strategies in the production and upgrading of bio-oils derived from lignocellulosic
biomass, J. Anal. Appl. Pyrol. 105 (2014) 55–74.
[5] M.S.A. Moraes, M.V. Migliorini, F.C. Damasceno, F. Georges, A. Almeida, C.A.
Zini, R.A. Jacques, E.B. Caramão, Qualitative analysis of bio oils of agricultural
residues obtained through pyrolysis using comprehensive two dimensional gas
chromatography with time-of-flight mass spectrometric detector, J. Anal. Appl. Pyrol.
98 (2012) 51–64.
[6] J. Caia, Y. He , X. Yu , S.W. Banks , Y. Yang ,X. Zhang, Y. Yu ,R. Liu ,A.V.
Bridgwater, Review of physicochemical properties and analytical characterization of
lignocellulosic biomass, Renew. Sust. Energ. Rev. 76 (2017) 309–322.
[7] S. Vaz Jr., Analytical techniques for the chemical analysis of plant biomass and
biomass products, Anal. Methods 6 (2014) 8094–8105.
[8] M.A. Carriquiry, X. Du, G.R. Timilsina, Second generation biofuels: economics and
policies, Energ. Policy 39 (2011) 4222–4234.

16
[9] J.S. Lim, Z.A. Manan, S.R.W. Alwi, H. Hashim, A review on utilization of biomass
from rice industry as a source of renewable energy, Renew. Sust. Energ. Rev. 16 (2012)
3084–3094.
[10] S. Xiu, A. Shahbazi, Bio-oil production and upgrading research: A review, Renew.
Sust. Energ. Rev. 16 (2012) 4406–4414.
[11] J. Álvarez-Flórez, E. Egusquiza, Analysis of damage caused by siloxanes in
stationary reciprocating internal combustion engines operating with landfill gas, Eng.
Fail. Anal. 50 (2015) 29-38.
[12] O. Sevimoglu, B. Tansel, Composition and source identification of deposits
forming in landfill gas (LFG) engines and effect of activated carbon treatment on
deposit composition, J. Environ. Manage. 128 (2013) 300-305.
[13] J.W. Wu, Y. Shi, Y.X. Zhu, Y.C. Wang, H.J. Gong, Mechanisms of Enhanced
Heavy Metal Tolerance in Plants by Silicon: A Review, Pedosphere 23 (2013) 815-825.
[14] Z. Song, H. Wang, P.J. Strong, Z. Li, P. Jiang, Plant impact on the coupled
terrestrial biogeochemical cycles of silicon and carbon: Implications for biogeochemical
carbon sequestration, Earth-Sci. Rev. 115 (2012) 319-331.
[15] Brazilian National Agency of Petroleum, Natural Gas and Biofuels. Resolution
ANP 52, DOU 30.12.2010.
[16] Ministry of Agriculture, Livestock and Food Supply, Normative Instruction nº 5,
DOU 23.02.2007.
[17] A.T. Duarte, A.R. Borges, A.V. Zmozinski, M.B. Dessuy, B. Welz, J.B. de
Andrade, M.G.R. Vale, Determination of lead in biomass and products of the pyrolysis
process by direct solid or liquid sample analysis using HR-CS GF AAS, Talanta 146
(2016) 166–174.
[18] A.T. Duarte, M.B. Dessuy, M.G.R. Vale, B. Welz, J.B. de Andrade, Sequential
determination of Cd and Cr in biomass samples and their ashes using high-resolution
continuum source graphite furnace atomic absorption spectrometry and direct solid
sample analysis, Talanta 115 (2013) 55–60.
[19] W.T. Tsai, S.C. Liu, C.H. Hsieh, Preparation and fuel properties of biochars from
the pyrolysis of exhausted coffee residue, J. Anal. Appl. Pyrol. 93 (2012) 63–67.
[20] J.S. Le Blond, S. Strekopytov, C. Unsworth, B.J. Williamson, Testing a new
method for quantifying Si in silica-rich biomass using HF in a closed vessel microwave
digestion system, Anal. Methods 3 (2011) 1752-1758.

17
[21] J.A. Tafur-Marinos, M. Ginepro, L. Pastero, V. Zelano, Digestion Procedures for
the Elemental Analysis of Wood by Inductively Coupled Plasma–Optical Emission
Spectrometry, Anal. Lett. 49 (2016) 1722-1733.
[22] S. Singh, L.C. Ram, R.E. Masto, S.K. Verma, A comparative evaluation of
minerals and trace elements in the ashes from lignite, coal refuse and biomass fired
power plants, Int. J. Coal Geol. 87 (2011) 112–120.
[23] M. Hasan, A. Salam, A.M.S. Alam, Identification and characterization of trace
metals in black solid materials deposited from biomass burning at the cooking stoves in
Bangladesh, Biomass Bioenerg. 33 (2009) 1376–1380.
[24] V.R.A. Filho, J.A.G. Neto, Evaluation of lubricating oil preparation procedures for
the determination of Al, Ba, Mo, Si and V by high-resolution continuum source FAAS,
Anal. Sci. 25 (2009) 95–100.
[25] Z. Kowalewska, J. Pilarczyk, Ł. Gościniak, Spectral aspects of the determination of
Si in organic and aqueous solutions using high-resolution continuum source or line
source flame atomic absorption spectrometry, Spectrochim. Acta B 120 (2016) 45–56.
[26] W. Boschetti, L.M.G. Dalagnol, M. Dullius, A.V. Zmozinski, E.M. Becker,
M.G.R. Vale, J.B. de Andrade, Determination of silicon in plant materials using direct
solid sample analysis with high-resolution continuum source graphite furnace atomic
absorption spectrometry, Microchem. J. 124 (2016) 380–385.
[27] L.C.C. de Oliveira, M.A. Vieira, A.S. Ribeiro, M.T. Lisboa, R.A. Gonçalves, R.C.
de Campos, Determination of silicon in vegetable oil and biodiesel by high-resolution
continuum source flame atomic absorption spectrometry using sample dilution with
xylene, Energ. Fuels 26 (2012) 7041−7044.
[28] R.S. Amais, G.L. Donati, D. Schiavo, J.A. Nóbrega, A simple dilute-and-shoot
procedure for Si determination in diesel and biodiesel by microwave-induced plasma
optical emission spectrometry, Microchem. J. 106 (2013) 318–322.
[29] M.S. Luz, P.V. Oliveira, Niobium carbide as permanent modifier for silicon
determination in petrochemical products by emulsion-based sampling GF AAS, Fuel
116 (2014) 255–260.
[30] M.F. Gazulla, M. Rodrigo, M. Orduña, M.J. Ventura, C. Andreu, High precision
measurement of silicon in naphthas by ICP-OES using isooctane as diluents, Talanta
164 (2017) 563–569.
[31] B. Welz, H. Becker-Ross, S. Florek, U. Heitmann, High-Resolution Continuum
Source AAS, Wiley-VCH, Weinheim, 2005.

18
[32] U. Heitmann, B. Welz, D.L.G. Borges, F.G. Lepri, Feasibility of peak volume, side
pixel and multiple peak registration in high-resolution continuum source atomic
absorption spectrometry, Spectrochim. Acta B 62 (2007) 1222–1230.
[33] M. Resano, E. Garcíıa-Ruiz, High-resolution continuum source graphite furnace
atomic absorption spectrometry: is it as good as it sounds? A critical review, Anal.
Bioanal. Chem. 399 (2011) 323–330.
[34] M. Resano, M. Aramendía, M.A. Belarra, High-resolution continuum source
graphite furnace atomic absorption spectrometry for direct analysis of solid samples and
complex materials: a tutorial review, J. Anal. Atom. Spectrom. 29 (2014) 2229–2250.
[35] U. Kurfürst, Solid Sample, Springer, Heidelberg, 1988.
[36] B. Welz, M. Sperling, Atomic Absorption Spectrometry, 3rd edn., Wiley-VCH,
Weinheim, 1999.
[37] W. Boschetti, R.T. Rampazzo, M.B. Dessuy, M.G.R. Vale, A.O. Rios, P. Hertz, V.
Manfroi, P.G. Celso, M.F. Ferrão, Detection of the origin of Brazilian wines based on
the determination of only four elements using high-resolution continuum source flame
AAS, Talanta 111 (2013) 147–155.
[38] E. Peña-Vázquez, M.C. Barciela-Alonso, C. Pita-Calvo, R. Domínguez-González,
P. Bermejo-Barrera, Use of high-resolution continuum source flame atomic absorption
spectrometry (HR-CS FAAS) for sequential multi-element determination of metals in
seawater and wastewater samples, J. Appl. Spectrosc. 82 (2015) 681–686.
[39] B. Gómez-Nieto, M.J. Gismera, M.T. Sevilla, J.R. Procopio, Fast sequential multi-
element determination of major and minor elements in environmental samples and
drinking waters by high-resolution continuum source flame atomic absorption
spectrometry, Anal. Chim. Acta 854 (2015) 13–19.
[40] B. Gómez-Nieto, M.J. Gismera, M.T. Sevilla, J.R. Procopio, Determination of
essential elements in beverages, herbal infusions and dietary supplements using a new
straightforward sequential approach based on flame atomic absorption spectrometry,
Food Chem. 219 (2017) 69–75.
[41] M.S.A. Moraes, F. Georges, S.R. Almeida, F.C. Damasceno, G.P.S. Maciel, C.A.
Zini, R.A. Jacques, E.B. Caramão, Analysis of products from pyrolysis of Brazilian
sugar cane straw, Fuel Process. Technol. 101 (2012) 35–43.
[42] W.P. Linak, J.O.L. Wendt, Trace metal transformation mechanisms during coal
combustion, Fuel Process. Technol. 39 (1994) 173–198.

19
Table 1. Graphite furnace heating program for Si determination by HR-CS SS-GF AAS
in different matrices. Argon was purged during all stages at 2.0 L min–1, except for
atomization, when it was stopped.
Stage Temperature (°C) Ramp (°C s–1) Hold (s)
Drying 1 90 10 20
Drying 2 200 10 50a, 30b
Pyrolysis 1200a, 1400b 300 30
Atomization 2650 3000 7
Cleaning 2650 0 5
a
Pyrolysis water and bio-oil, bbiomass and ash

20
Table 2. Temperature program for SS platform coating with Rh as permanent modifier;
the program cycle was repeated 9 times up to stage 4, and the final injection (total of 10)
has run all temperature stages.
Stage Temperature (°C) Ramp (°C s–1) Hold (s)
1 100 10 20
2 150 30 30
3 250 20 35
4 1000 1000 10
5 2000 200 5

21
Table 3. Temperature program of the microwave-assisted acid digestion (HNO3/HF) of
biomass and ash samples with 1000 W, except for the cooling step.
Temperature (°C) Ramp (min) Hold (min)
100 20 5
140 20 5
160 15 5
180 15 15
20 0 40

22
Table 4. Figures of merit for the determination of Si in pyrolysis products and biomass
using HR-CS AAS.
Parameter HR-CS F AAS HR-CS GF AAS
Wavelength (nm) 251.611 251.611 221.174
LOD (mg kg–1) 6a 0.09b 3c
LOQ (mg kg–1) 20a 0.3b 10c
c0 (mg L–1) 2 n/a n/a
m0 (ng) n/a 0.3 2
n/a: not applicable, calculated for a sample mass of a100 mg, b10 mg, and c0.1 mg.

23
Table 5. Evaluation of the accuracy of the determination of Si in biomass, bio-oil, and
ash using HR-CS AAS. The uncertainties are expressed as 95% confidence interval (n =
5).
Si concentration (% m m–1)
CRM
Certified HR-CS F AAS HR-CS GF AAS
Bush branches and leaves
0.60 ± 0.07 0.60 ± 0.04* 0.62 ± 0.07**
(NSC DC 73349)
Wear-metals in lubricating
103  10–4 n/a (112 ± 10)  10–4*
oil (SRM 1084a)
Marine sediment (BCSS-1) 30.9 ± 0.5 32.2 ± 1.8* 32.1 ± 3.1**
n/a: not applicable, *Determination at 251.611 nm, **Determination at 221.174 nm

24
Table 6. Summary of the methods used throughout this study for each sample.
Matrix Sample preparation Technique
Bio-oil Dilution with propan-1-ol HR-CS GF AAS
(251.611 nm)
Pyrolysis water Direct analysis HR-CS GF AAS
(251.611 nm)
Biomass and ash (coconut Microwave-assisted acid digestion HR-CS F AAS
and palm tree fiber) with HNO3 and HF. (251.611 nm)
Biomass and ash (peach Direct analysis HR-CS SS-GF AAS
pit) (221.174 nm)

25
Table 7. Si determination in pyrolysis fractions of coconut fiber, palm tree fiber, and
peach pit using HR-CS AAS. The uncertainties are expressed as 95% confidence
interval (n = 5). Si loss was obtained by the difference between the biomass (considered
100%) and the sum of the three pyrolysis products. n/a: not applicable.
Material Si concentration Mass fraction
Sample Material
mass (g) (mg kg–1) Si (%)
Biomass 100 (1.16 ± 0.40)  103 100
Bio-oil 13.2 1.38 ± 0.45 0.016
Coconut fiber Pyrolysis water 29.4 13.4 ± 0.7 0.34
Ash 29.2 (4.11 ± 1.12)  103 103
Loss n/a n/a ~0
Biomass 97.7 (9.15 ± 1.59)  103 100
Bio-oil 8.5 1.06 ± 0.27 0.0010
Palm tree fiber Pyrolysis water 29.1 3.31 ± 0.79 0.011
Ash 27.8 (13.7 ± 2.7)  103 43
Loss n/a n/a 57
Biomass 260 (3.46 ± 0.14)  103 100
Bio-oil 21.5 1.70 ± 0.51 0.0041
Peach pit Pyrolysis water 88.9 12.1 ± 2.2 0.12
Ash 68.5 (12.2 ± 1.0)  103 93
Loss n/a n/a 6.9

26
Fig. 1. Time-resolved absorbance spectrum of a 60 mg L–1 Si solution in the vicinity of
251.611 nm using HR-CS F AAS with the SFS6 valve. Injection time: 0.5 s.

Fig. 2. Pyrolysis and atomization curves for Si determination using 300 μg Rh + 10 μg

Pd + 6 μg Mg as modifier for () 50 ng Si as aqueous standard solution, () peach pits
biomass, and (▲) ash. Aint values for biomass and ash samples are normalized to a mass
of 0.1 mg. Wavelength: 221.174 nm; Tpyr = 1400 °C and Tatom = 2650 °C. Error bars
correspond to the standard deviation (n = 5).

27
Fig. 3. Time-resolved absorbance spectra of A) 6 mg of diluted bio-oil and B) pyrolysis
water (10 L) of peach pit at the vicinity of 251.611 nm, C) biomass and D) ash of
peach pit were analyzed, 0.1 mg of each at 221.174 nm, using HR-CS GF AAS.
Chemical modifiers 300 g Rh + 10 g Pd + 6 g Mg; Tpyr = 1400 °C and Tatom = 2650
°C.

28
Fig. 4. Relation between the solid sample mass, () ash and () biomass, and the
integrated absorbance. Wavelength: 221.174 nm; chemical modifiers 300 g Rh + 10
g Pd + 6 g Mg; Tpyr = 1400 °C and Tatom = 2650 °C.

29
30
Highlights:

 Palm tree fiber, coconut fiber and peach pit were submitted to a pyrolysis
process.
 A mass balance of Si has been established in all products of the pyrolysis
process.
 The bio-oil contained <2 mg kg-1 Si or <0.02% of the total Si in biomass
samples.
 The Si content in the bio-oil is favorable for its use as an alternate fuel.

31