Fuel 129 (2014) 314–317

Contents lists available at ScienceDirect

Fuel
journal homepage: www.elsevier.com/locate/fuel

Short communication

Determination of chlorine in solid fuels using an improved Eschka

method
Muhammad Usman Rahim, Xiangpeng Gao, Hongwei Wu ⇑
School of Chemical and Petroleum Engineering, Curtin University, GPO Box U1987, Perth, WA 6845, Australia

h i g h l i g h t s

 Eschka method underestimates the Cl contents in low-rank solid fuels such as biomass.
 Cl can escape with flue gas during the Eschka combustion of biomass.
 An improved Eschka method is developed to recover the escaped Cl for quantification.
 The improved method has small error and low limit of quantification.

a r t i c l e i n f o a b s t r a c t

Article history: Conventional Eschka method is a standard method for the determination of chlorine (Cl) in solid fuels.
Received 18 February 2014 However, the method significantly underestimates the Cl contents in some low-rank solid fuels (e.g., bio-
Received in revised form 25 March 2014 mass) due to the escape of Cl with flue gas from the Eschka combustion process. An improved Eschka
Accepted 27 March 2014
method is developed to recover the escaped Cl in the flue gas via converting it into HCl(g) by flue gas
Available online 13 April 2014
re-combustion for capture and quantification. The improved Eschka method can accurately quantify Cl
in a range of solid fuels studied (i.e., a mallee bark, a Victorian brown coal, a sub-bituminous coal, an
Keywords:
anthracite coal, and a polyvinyl chloride), with a small relative standard error (±2%) and a low limit of
Biomass
Coal
quantification (0.015 mg/g fuel, dry basis) under present experimental conditions.
Chlorine Ó 2014 Elsevier Ltd. All rights reserved.
Eschka method
Combustion

1. Introduction significant errors due to issues including incomplete combustion

Thermochemical utilisation of chlorine (Cl)-containing solid
fuels (e.g., coal [1–3], biomass [4–6], waste plastic [7,8] and refused
derived fuel [9,10]) contributes significantly to the emissions of
hydrogen chloride [11], chloromethane [12], polychlorinated
dibenzo-p-dioxins and dibenzofurans (PCDDs/Fs) [9,13,14] and
particulate matter [15]. Such emissions cause various environmen-
tal issues such as acid rain [16] and stratospheric ozone depletion
[12]. Accurate quantification of Cl in these fuels is essential to
understand Cl transformation during thermochemical processes
and mitigate related environmental issues.
Conventional combustion-based methods such as bomb com-
bustion [17], high-temperature combustion [18] and Eschka
method [19] are commonly used for quantifying Cl in solid fuels.
Particularly, the Eschka method is often used as a standard method
for benchmarking [17]. However, these methods may lead to
⇑ Corresponding author. Tel.: +61 8 92667592; fax: +61 8 92662681.
E-mail address: h.wu@curtin.edu.au (H. Wu).
[20] (for bomb combustion), loss of chlorine [21] (for high-temper-
ature combustion method) and/or the escape of Cl (for Eschka)
[22]. To address these shortcomings, a two-step method was
developed [23], which employed slow pyrolysis at a suitable tem-
perature to separate fuel-Cl into volatile-Cl (i.e., Cl released into
volatiles) and char-Cl (i.e., Cl retained in char). The volatile-Cl is
quantified by converting it into HCl(g) via combustion, capturing
the converted HCl via sodium hydroxide solution, and analysing
the captured Cl via ion chromatography (IC). The char-Cl was quan-
tified via the conventional Eschka method. While the two-step
method can accurately quantify the Cl content in solid fuels, there
are three disadvantages associated with the method: (a) a suitable
pyrolysis temperature may need to be tested to ensure the
stabilization of Cl in char and hence the complete capture of Cl
by Eschka mixture (sodium carbonate and magnesium oxide); (b)
the analytical time is prolonged (over 8 h), which leads to low effi-
ciency; and (c) the analysis of char-Cl requires additional collec-
tion, transfer and weighing of char samples in the pyrolysis
reactor, leading to increased analytical errors.

http://dx.doi.org/10.1016/j.fuel.2014.03.070
0016-2361/Ó 2014 Elsevier Ltd. All rights reserved.
 M.U. Rahim et al. / Fuel 129 (2014) 314–317
Therefore, this study aims to develop an improved Eschka
method for accurate and time-efficient quantification of Cl in var-
ious solid fuels. It is built on the conventional Eschka method but
includes the combustion of the flue gas from the Eschka combus-
tion process, converting the escaped Cl into HCl(g), followed by
subsequent capture of the escaped Cl using 0.1 M NaOH solution.
The streamlined method enables accurate quantification of Cl in
solid fuels, with small analytical errors and low limit of quantifica-
tion at a reasonable analytical time.
2. Materials and methods
2.1. Materials
Five solid fuel samples, including a mallee bark, three coals of
different ranks, and a polyvinyl chloride (PVC), were studied to
cover broad fuel types and ranks as well as different occurrence
forms of Cl (i.e., both inorganic and organic). The malee bark was
separated from mallee trees harvested in Narrogin, Western Aus-
tralia. The three coals included a Victorian brown coal, a sub-bitu-
minous coal purchased from Alpha Resources (AR-2780) with a
standard Cl content of 1.3 mg/g fuel (dry basis, as reported by
the supplier), and an anthracite coal that the sample was available
in the laboratory. The PVC was purchased from sigma Aldrich
(Product No. 81388). All the samples were sieved to 75–150 lm ,
air-dried at 40 °C, sealed in plastic bags, and stored in freezer at -
4 °C prior to experiments.
2.2. Improved Eschka method
Fig. 1 presents a schematic diagram of the experimental setup
employed for the improved Eschka method. It consists of an Eschka
reactor required for the conventional Eschka combustion process
and an additional combustion reactor for flue gas re-combustion,
which is connected to the Eschka reactor via a quartz tube. During
an experiment, 150 mg fuel sample was mixed with 300 mg of
Eschka flux (a mixture of sodium carbonate and magnesium oxide,
purchased from Sigma Aldrich, with a product number of 00166) in
a platinum (Pt) boat, which was further covered with 200 mg of
Eschka mixture. The flue gas re-combustion reactor was preheated
to 950 °C and the quartz connection tube was preheated and main-
tained at 675 °C. The sample-loaded Pt boat was then placed in the
Eschka reactor, which was heated to 675 °C (heating rate: 10 °C/
min, holding time: 2 h; air flow rate: 2 L/min) in accordance with
the Eschka method [19]. The flue gas was continuously swept out
the Eschka reactor and burned in the combustion reactor in the
presence of two streams of oxygen, i.e., primary oxygen (flow rate:
0.4 L/min) and secondary oxygen (flow rate: 0.4 L/min). The flue
gas re-combustion process converted the escaped Cl into HCl(g),
which was subsequently captured by 0.1 M NaOH solution (in form
of chlorides) through gas bubblers. The chlorides in the NaOH solu-
tion were quantified by an IC system (Dionex ICS-1100), with an
IonPac AS22 fast analytical column (4  150 mm) and 2.25 mM
Fig. 1. Schematic diagram of the experimental setup for the improved Eschka method.
315
Na2CO3/0.7 mM NaHCO3 solution as an eluent. The residue in the
Pt boat was dissolved directly in ultra-pure water for chlorides
analysis using the same IC system.
2.3. Sample analysis
The actual Cl contents of the fuels were quantified using a two-
step method detailed elsewhere [23] for benchmarking purpose.
Proximate analysis was carried out using a Mettler thermogravi-
metric analyzer (TGA) according to ASTM E870-82 [24]. Ultimate
analysis, including Carbon (C), Hydrogen (H), Nitrogen (N) and Sul-
phur (S), was conducted using a CHNS/O elemental analyzer
(model: Perkin–Elmer 2400 series II), following Australian Stan-
dard AS1038.6.4 [25]. Oxygen (O) content was calculated by differ-
ence on dry and ash-free (daf) basis. The proximate and ultimate
analyses of the fuels are given in Table S1 in the Supplementary
Material. The ignition temperatures of selected fuels were deter-
mined via a method detailed elsewhere [26].
3. Results and discussion
3.1. Drawbacks of conventional Eschka method
Fig. 2a presents the Cl contents of the five fuels determined by
the conventional Eschka method, benchmarking with the corre-
sponding actual Cl contents determined by the two-step method.
The two-step method has been validated in our previous paper
[23] via a series of standard samples (i.e., NaCl-loaded cellulose
chars and PVC) with known Cl contents. The Cl contents deter-
mined by the conventional Eschka method are in agreement with
the actual values for the three coal samples. However, the conven-
tional Eschka method significantly underestimates the Cl contents
in the mallee bark and the PVC, due to at least two possible rea-
sons. One is the escaped Cl that is not captured by the Eschka flux.
The other reason is the incomplete recovery of Cl in the mixture of
fuel ash and Eschka flux. However, the latter is less likely because
the conventional Eschka method is capable of quantifying all Cl in
the sample via the Eschka flux if the Cl is released as HCl(g) [27].
Therefore, the results clearly suggest that at least a portion of Cl
in the fuel escaped from the mixture during the Eschka combustion
process is released in forms other than HCl(g). This leads to the
underestimation of Cl when the conventional Eschka method is
used.
A brief literature review was then undertaken on possible forms
of Cl released during the pyrolysis and combustion of solid fuels
[1,4,12,21,28–39], as summarized in Table S2 of the Supplemen-
tary Material. Depending on fuel types and temperature–time his-
tory experienced by fuel particles during the pyrolysis and
combustion, Cl can be released from the reacting fuel particles in
various forms. The possible forms are HCl(g) [1,4,21,29–39], chlo-
romethane(CH3Cl) [12], Cl organically bound with tar (tar-Cl)
[21,34] and gaseous potassium chloride KCl(g) [4,28–30,36,37].
This provides some plausible explanations on the applicability of
316 M.U. Rahim et al. / Fuel 129 (2014) 314–317

Eschka method (mg/g fuel, dry basis)

600 600
(a) actural chlorine content (b)

Chlorine content by improved

(mg/g fuel, dry basis)
550 550
quantified via conventional

Chlorine content
500 Eschka method 500 PVC
A: anthracite coal
450 B: Victorian brown coal 450 Victorian
C: sub-bituminous coal brown coal
6 6 mallee bark
4 D: mallee bark 4
E: PVC sub-bituminous coal
2 2
anthracite coal
0 0
A B C D E 0 2 4 6 450 500 550 600
Actual chlorine content in fuel
(mg/g fuel, dry basis)

600 3
(c) chlorine in flue gas (d) A: with flue gas combustion
(mg/g fuel, dry basis)

550 chlorine in solid residue

(mg/g fuel, dry basis)

B: without flue gas combustion

Chlorine in flue gas

Chlorine content

500 2 C: combustion of flue gas

A: anthracite coal after condensation
450 B: Victorian brown coal at 110 °C
C: sub-bituminous coal
6 1
D: mallee bark
4 E: PVC
2
0 0
A B C D E A B C

Fig. 2. Quantification of Cl in the anthracite coal, Victorian brown coal, sub-bituminous coal, mallee bark, and PVC samples. Panel (a): comparison of Cl contents determined
via conventional Eschka method and the actual Cl contents in the fuels; Panel (b): benchmarking of Cl contents determined via the improved Eschka method against the
actual Cl contents in the fuels; Panel (c): distribution of Cl in solid residue and flue gas, as quantified via the improved Eschka method; and Panel (d): recovery of Cl in flue gas
during the Eschka combustion of the mallee bark. The actual Cl contents in the fuels were determined by the two-step method that was detailed elsewhere [23].

the conventional Eschka method to some samples. For example,
Table S2 (see Supplementary Material) shows that HCl(g), which
can be completely captured by the Eschka flux, is the main form
of Cl released during the pyrolysis and combustion of coals. There-
fore, the conventional Eschka method is capable of quantifying Cl
in the three coals studied.
The information in Table S2 (see Supplementary Material) also
provides some hints on the possible forms of Cl escaped during
the combustion of the mixture of Eschka flux and the mallee bark
(or the PVC). Apart from HCl(g), Cl can also be released as chloro-
methane (CH3Cl), Cl organically bound with tar, and/or KCl(g) dur-
ing the pyrolysis or combustion of biomass. However, it is unlikely
for Cl to be released in the form of KCl(g) from the reacting Eschka
flux because the Eschka combustion temperature (675 °C) is lower
than the temperature (>700 °C, see Table S2 in the Supplementary
Material) at which KCl(g) starts to release. On the other hand, the
release of CH3Cl and/or tar-Cl is most likely the cause of the under-
estimation of Cl content in the mallee bark via the conventional
Eschka method. This is supported by the reported temperature
range of 150–350 °C at which CH3Cl release takes place during bio-
mass pyrolysis [12]. The ignition temperature of the mallee bark
determined by the TGA is 260 °C that is also within the temper-
ature range. It appears that at least part of CH3Cl, of which the Cl
cannot be captured by Eschka flux, would have been released
before the ignition of the mallee bark. Similarly, part of Cl will be
released into tar during PVC pyrolysis at 260–300 °C (see
Table S2 in the Supplementary Material), leading to the underesti-
mation of Cl content in the PVC via the conventional Eschka
method because the Cl in tar-Cl cannot be captured by Eschka flux.
3.2. An improved Eschka method
The results so far suggest that the underestimation of Cl con-
tents in the mallee bark and the PVC via the conventional Eschka
method is mainly caused by the release (or escape) of Cl in forms
other than HCl(g). Therefore, to accurately quantify Cl in such fuels
based on the conventional Eschka method, the key is to modify the
method to successfully re-capture and quantify the escaped Cl in
flue gas. Therefore, as shown in Fig. 1, this study develops an
improved Eschka method that appends an additional combustion
reactor to the existing Eschka reactor. The purpose of the combus-
tion reactor is to re-combust all the flue gas released during the
Eschka combustion process and to convert all Cl in the flue gas into
HCl(g). The converted Cl is subsequently captured by the 0.1 M
NaOH solution, quantified and considered in determining the total
Cl in the fuel.
Fig. 2b presents the Cl contents of the five fuels determined by
the improved Eschka method, benchmarking with the actual Cl
contents determined by the two-step method [23]. The results
demonstrate the excellent agreement between the Cl contents in
all the fuels determined by the two methods. The results show that
the improved Eschka method is capable of accurately quantifying
Cl over a wide range (0.3–567 mg/g fuel, dry basis) in various
solid fuels such as biomass, coal, and plastics (mainly PVC). For
the improved Eschka method, the estimated relative standard error
is ±2% and the limit of quantification is 0.015 mg/g fuel (dry basis)
under the current sample size (150 mg) and reactor configura-
tions. Fig. 2c further presents the distribution of Cl in solid residue
and flue gas. While the Cl in flue gas is negligible for the three
coals, the data show that 48% and 10% of Cl are presented in flue
gas for the mallee bark and the PVC, respectively. This in turn
explains why the conventional Eschka method significantly under-
estimates the Cl content in the two fuels. Most importantly, the
results also demonstrate that such underestimation of Cl contents
is indeed due to the release of Cl in forms other than HCl(g). The
complete recovery of the released Cl through its conversion into
HCl(g) by flue gas re-combustion enables the improved Eschka
method to accurately quantify Cl in these fuels.
Further efforts were made to identify the forms of Cl released
with the flue gas during the combustion of the mixture of Eschka
flux and the mallee bark at 675 °C. Experiments were designed to
capture and collect the Cl released with the flue gas using the
0.1 M NaOH solution in three different modes: (1) with flue gas
in situ re-combustion; (2) without flue gas re-combustion; and
 M.U. Rahim et al. / Fuel 129 (2014) 314–317
(3) with the re-combustion of flue gas after the condensation of the
flue gas at 110 °C. The first mode is capable of quantifying total Cl
released with flue gas while the second mode only collects the
HCl(g) escaped from the Eschka flux (if any). In the third mode,
the temperature of the connection tube between the Eschka reac-
tor and the flue gas combustion reactor was set at 110 °C. The tem-
perature is sufficiently low to allow the condensation of heavy
unburned organic compounds (if any) but also high enough to
avoid moisture condensation that can trap HCl(g) and CH3Cl(g) in
the flue gas. Therefore, only light unburned organic compounds
and/or CH3Cl(g) in the flue gas will be re-combusted and quanti-
fied. The results are presented in Fig. 2d, which leads to two impor-
tant findings. One is that the Eschka flux is capable of capturing all
the Cl released as HCl(g) during the combustion of the mixture of
Eschka flux and the mallee bark. This is supported by the fact that
negligible Cl in the flue gas is collected without flue gas combus-
tion (see Fig. 2d). It should be noted that in the second mode, the
flue gas was kept at 500 °C before entering the gas bubblers to
avoid any HCl(g) loss as a result of moisture condensation. The
other finding is that the majority (60%) of the Cl in the flue gas
is recovered after condensing the flue gas at 110 °C, indicating that
the Cl escaped with the flue gas is not only in association with
heavy unburned organic compounds but also in the forms of light
organic compounds and/or gases (e.g., CH3Cl [12]).
Compared to the two-step method [23], the improved Eschka
method possesses at least two advantages because of the omission
of the lengthy pyrolysis step, although both methods are capable of
accurately quantifying Cl in various solid fuels. One is that the ana-
lytical time is shortened from over 8 h for the two-step method to
around 3 h for the improved Eschka method. The other advantage
is that the relative standard error of the improved Eschka method
(±2%) is smaller than that of the two-step method (±5%).
4. Conclusions
This study reports an improved Eschka method for recovering
the escaped Cl in the flue gas from the conventional Eschka method
via converting it into HCl(g) by flue gas re-combustion for subse-
quent capture and quantification. The improved method is applica-
ble to various solid fuels. Under the measurement configurations,
the improved method has a small relative standard error (±2%)
and a low limit of quantification (0.015 mg/g fuel, dry basis).
Acknowledgments
MUR acknowledges Curtin University, the University of the
Punjab and HEC Faculty Development Program of Pakistan Govern-
ment for the partial support to his PhD study.
Appendix A. Supplementary material
Fuel properties of the solid fuels studied (Table S1) and a brief
literature review on the possible forms of Cl released during
317
the pyrolysis or combustion of solid fuels including biomass, coal
and plastics (Table S2), downloadable at http://www.sciencedirect.
com. Supplementary data associated with this article can be
found, in the online version, at http://dx.doi.org/10.1016/j.fuel.
2014.03.070.
References
[1] Pearce WC, Hill JWF. Prog Energy Combust Sci 1986;12:117–62.
[2] Sui J, Xu M, Qiu J, Qiao Y, Yu Y, Liu X, Gao X. Energy Fuels 2005;19:1536–41.
[3] Tsubouchi N, Saito T, Ohtaka N, Nakazato Y, Ohtsuka Y. Energy Fuels
2013;27:5076–82.
[4] Björkman E, Strömberg B. Energy Fuels 1997;11:1026–32.
[5] Jenkins BM, Baxter LL, Miles Jr TR, Miles TR. Fuel Process Technol
1998;54:17–46.
[6] Bläsing M, Zini M, Müller M. Energy Fuels 2013;27:1439–45.
[7] Gui B, Qiao Y, Wan D, Liu S, Han Z, Yao H, et al. Proc Combust Inst
2013;34:2321–9.
[8] Jordan KJ, Suib SL, Koberstein JT. J Phys Chem B 2001;105:3174–81.
[9] Hatanaka T, Kitajima A, Takeuchi M. Environ Sci Technol 2005;39:9452–6.
[10] Vassilev SV, Braekman-Danheux C. Fuel Process Technol 1999;59:135–61.
[11] McCulloch A, Aucott ML, Benkovitz CM, Graedel TE, Kleiman G, Midgley PM,
et al. J Geohys Res D 1999;104:8391–403.
[12] Hamilton JTG, McRoberts WC, Keppler F, Kalin RM, Harper DB. Science
2003;301:206–9.
[13] Buekens A, Huang H. J Hazard Mater 1998;62:1–33.
[14] Grandesso E, Gullett B, Touati A, Tabor D. Environ Sci Technol
2011;45:3887–94.
[15] Gao X, Wu H. Energy Fuels 2011;25:4172–81.
[16] Lightowlers PJ, Cape JN. Atmos Environ 1988;22:7–15.
[17] Moszynski ZKA. J Appl Chem 1955;5:467–70.
[18] Australian standard, AS 1038.8.2, Part 8. 2: chlorine—high-temperature
combustion method, Standards Australia, 2003.
[19] Australian standard, AS 1038.8.1, Part 8. 1: chlorine—Eschka method,
Standards Australia, 1999.
[20] Ma W, Hoffmann G, Schirmer M, Chen G, Rotter VS. J Hazard Mater
2010;178:489–98.
[21] Rahim MU, Gao X, Garcia-Perez M, Li Y, Wu H. Energy Fuels 2012;27:310–7.
[22] Watanabe N, Tanikawa N, Oikawa T, Inoue S, Fukuyama J. J Mater Cycles Waste
Manag 2003;5:0069–76.
[23] Rahim MU, Gao X, Wu H. Energy Fuels 2013;27:6992–9.
[24] Standard test methods for analysis of wood fuels, E870–82 (reapproved 2006),
ASTM—American Society for Testing and Materials, 2006.
[25] Australian standard, AS1038.6.4, Part 6. 4: higher rank coal and coke—ultimate
analysis—carbon, hydrogen and nitrogen—instrumental method, Standards
Australia, 2005.
[26] Wang C, Wang F, Yang Q, Liang R. Biomass Bioenergy 2009;33:50–6.
[27] Cox JA. In: Stringer J, Banerjee DD, editors. Chlorine in coal, coal science and
technology, Amsterdam. Amsterdam: Elsevier Science Publishers; 1991.
[28] Knudsen JN, Jensen PA, Dam-Johansen K. Energy Fuels 2004;18:1385–99.
[29] Jensen PA, Frandsen FJ, Dam-Johansen K, Sander B. Energy Fuels
2000;14:1280–5.
[30] Tchoffor PA, Davidsson KO, Thunman H. Energy Fuels 2013;27:7510–20.
[31] Quyn DM, Wu H, Li C-Z. Fuel 2002;81:143–9.
[32] Takeda M, Ueda A, Hashimoto H, Yamada T, Suzuki N, Sato M, et al. Fuel
2006;85:235–42.
[33] Shao D, Hutchinson EJ, Cao H, Pan W-P, Chou C-L. Energy Fuels
1994;8:399–401.
[34] Ma S, Lu J, Gao J. Energy Fuels 2002;16:338–42.
[35] Miranda R, Pakdel H, Roy C, Vasile C. Polym Degrad Stabil 2001;73:47–67.
[36] van Lith SC, Alonso-Ramírez V, Jensen PA, Frandsen FJ, Glarborg P. Energy Fuels
2006;20:964–78.
[37] van Lith SC, Jensen PA, Frandsen FJ, Glarborg P. Energy Fuels
2008;22:1598–609.
[38] Zhang C, Wang Y, Yang Z, Xu M. Fuel 2006;85:2034–40.
[39] Zheng X-G, Tang L-H, Zhang N, Gao Q-H, Zhang C-F, Zhu Z-B. Energy Fuels
2003;17:896–900.