Design and thermal characterization of an

induction-heated reactor for pyrolysis of solid waste

Oscar Sosa Sabogal, Sylvie Valin, Sébastien Thiery, Sylvain Salvador

To cite this version:

Oscar Sosa Sabogal, Sylvie Valin, Sébastien Thiery, Sylvain Salvador. Design and thermal characteri-
zation of an induction-heated reactor for pyrolysis of solid waste. Chemical Engineering Research and
Design, 2021, 173, pp.206-214. �10.1016/j.cherd.2021.07.018�. �hal-03295902�

HAL Id: hal-03295902

https://imt-mines-albi.hal.science/hal-03295902
Submitted on 8 Sep 2021

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est

archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents
entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non,
lished or not. The documents may come from émanant des établissements d’enseignement et de
teaching and research institutions in France or recherche français ou étrangers, des laboratoires
abroad, or from public or private research centers. publics ou privés.
Design and thermal characterization of an
induction-heated reactor for pyrolysis of solid
waste

Oscar Sosa Sabogal a,b,∗ , Sylvie Valin a , Sébastien Thiery a ,

Sylvain Salvador b
a Univ. Grenoble Alpes, CEA, LITEN, DTBH, 17 rue des Martyrs, Grenoble Cedex 09, 38000 France
b Centre RAPSODEE, IMT Mines Albi, CNRS UMR 5302, Campus Jarlard, Albi Cedex 09, 81013, France

a b s t r a c t

A small-scale induction heated reactor (IHR) was specifically developed to study fast pyrol-
ysis, here investigated as the first step of gasification process, representing some of the
reaction conditions encountered in a fluidized bed reactor. First, the thermal response of
the system was characterized at transient and steady state, and CFD calculations were per-
formed to have a complete description of the temperature profiles inside the reactor. The
novel device can handle a few grams of solid at temperatures up to 900 ◦ C, allowing high
Keywords: heating rates (near 80 ◦ C/s) and a uniform distribution of temperature in the sample. Sec-
Pyrolysis ondly, the pyrolysis of a solid recovered fuel (SRF) sample was carried out at 800 ◦ C, and
Gasification the distribution and composition of reaction products were analyzed and compared with
Solid waste tests performed in a pilot scale fluidized bed reactor (FBR). The results obtained in the IHR
Fluidized bed reactor showed a good reproducibility. The same main gas and tar species were measured in the
Induction heating IHR and FBR, with some differences in gas and tar composition that were attributed to the
extent of secondary reactions, enhanced by higher heat transfer rates and the presence of
bed material in the FBR.

1. Introduction The produced gas can be directly burnt to produce energy and heat
or, after cleaning, used in a synthesis process to produce biofuels or
Solid waste that cannot be reused or recycled can be transformed chemicals. The yields of products and their composition highly depend
through energy recovery processes, an interesting alternative to land- on reactor design, process parameters, and feedstock characteristics.
fill. Waste derived fuels are produced from industrial or household Among the available technologies in thermochemical conversion, flu-
waste, which include common materials with high calorific values like idized bed reactors are widely used in the pyro-gasification of biomass
plastics, textiles, wood, and elastomers (Garcés et al., 2016). After incin- and waste since they are suitable for heterogeneous streams (Benedikt
eration, pyrolysis and gasification are the most common valorization et al., 2018). Compared to fixed bed reactors, they are characterized
routes. Pyrolysis is a thermochemical process where the carbonaceous by higher mass and energy transfer rates, achieving higher conversion
feedstock decomposes in absence of oxygen to produce permanent rates and low tar yields (Park et al., 2018).
volatiles (gas), condensable hydrocarbons (tar) and often a solid residue Testing different feedstocks or conditions in a fluidized bed pilot can
(char). Gasification usually occurs at temperatures above 700 ◦ C, in be costly in time and resources. Lab scale devices offer a simpler and
presence of an oxidizing agent that is fed below the stoichiometric more flexible solution, however, the reaction conditions found in an
amount needed for total oxidation. industrial fluidized bed reactor are difficult to scale down (Leion et al.,
2018). The rate and extent of decomposition of the feedstock during
the pyrolysis process, is mostly influenced by the undergone thermal
history. This history is classically defined by three parameters: the heat-
ing rate, the final temperature, and the total residence time (Pasel and
∗
Corresponding author. Wanzl, 2003). Slow heating rates induce high residence times and favor
char formation, while high heating rates and high temperatures favor These characteristics include heating rates between 100 and
gas products (Efika et al., 2018). 1000 ◦ C/s (Nilsson et al., 2012), a temperature range between
Most researchers have used thermogravimetric analysis (TGA) to 750 and 900 ◦ C, and good gas/solid heat and mass transfer
study the pyrolysis of solid carbonaceous feedstocks and char gasifica- conditions. Gas residence times inside the hot reactor are usu-
tion kinetics (Aluri et al., 2018; Porshnov et al., 2018; Zhou et al., 2015). In
ally between 0.5 and 10 s for bubbling fluidized bed reactors,
this technique, mass loss is continuously measured while the sample
and between 0.5 and 1 s for circulating fluidized bed reac-
is subjected to a temperature treatment at a constant heating rate or a
tors (Marshall et al., 2014). For the present device, a residence
constant temperature under controlled atmosphere. However, conven-
tional thermogravimetric analyzers present some limitations such as time between 1 and 5 s was targeted, long enough to enable
low heating rates (∼1 ◦ C/s) and limited gas-solid contact (Saadatkhah significant secondary reactions of volatiles (H2 , CO, CO2 , H2 O
et al., 2020; Samih and Chaouki, 2015). Sample mass is around 10–100 and hydrocarbons, including tar species). These reactions take
mg, which makes it difficult to obtain representative results, especially place at temperatures above 600 ◦ C before the volatiles exit the
for mixtures as heterogeneous as waste derived fuels (Robinson et al., reactor freeboard zone (Barr et al., 2019).
2016). Curie-point and wire mesh reactors achieve very high heating The reactor was designed to handle a few grams of sample
rates (>104 ◦ C/s), however sample mass is also in the scale of a few and to minimize heat and mass transfer limitations. A homo-
milligrams and reaction products are difficult to quantify (Zhu et al.,
geneous temperature in the sample was desired along the
2020). Other devices use larger amounts of sample, like fixed bed reac-
whole reaction time. Temperature gradients within the reac-
tors (Hwang et al., 2014), drop tube and boat furnaces (Daouk et al.,
tor were reduced by preheating the carrier/gasification agent
2018). Heat and mass transfer limitations may become significant if the
heating rate is not high enough, if large particles are used or if there is
gas, before entering into contact with the sample. Finally,
not a good contact between the gas and the solid (Cortazar et al., 2020). another objective was to collect all the reaction products
Some of the pyrolysis and gasification steps are endothermic, which (solids, condensable and permanent gases) for further anal-
means that enough heat must be provided to the reactor to sustain yses and quantification.
these reactions. In fluidized bed reactors, heat can be transferred from
the bed material (heated in a separate vessel), from a hot fluidizing
gas, by the oxidation of a feedstock fraction (autothermal operation) or 2.2. Detailed description of the setup
by external means (Zhang et al., 2018). In lab scale installations heat
is transferred to the reactor wall externally. Electric resistance heat-
The designed and developed lab-scale setup is shown in Fig. 1.
ing is one of the most common sources at this scale. It offers a good
It consists of an induction-heated reactor, followed by a con-
temperature control; however, it usually takes long time to reach the
densable and gas products collection system, and analysis
setpoint temperature due to thermal inertia. High temperature differ-
ences appear between the reactor wall and the sample, therefore, a modules.
major response delay is observed (Latifi and Chaouki, 2015). Alternative The reactor consists of a stainless-steel tube of 560 mm in
sources that can provide faster and more uniform heating are worth height, 31.75 mm in external diameter and with a thickness
considering. of 0.8 mm. The type of stainless steel used for the reactor was
Widely used in the treatment of metal materials, induction heating 316L; it can withstand temperatures up to 900 ◦ C. A constant
is a contactless method in which an AC power source is used to sup- nitrogen flow fed from the bottom of the reactor is used to
ply an alternating current to a coil, which is wrapped around the work purge the system and to sweep the produced volatiles during
piece. This current generates an electromagnetic field on the work piece
the course of the experiment. Flowrate is set with a BROOKS’s
and consequently heats it by two phenomena: creation of eddy currents
5851S mass flowmeter controlled by a LabVIEW interface.
and magnetic hysteresis (Latifi and Chaouki, 2015). Induction heating
A custom build heat exchanger helps to preheat the enter-
offers several key features like rapid heating rates, precise tempera-
ture control and high energy efficiency, which makes it an interesting
ing gas. The arrangement consists of two concentric metallic
energy source for the study of thermal conversion processes (Henkel, Inconel 600 tubes, with a height of 50 mm, a thickness of 0.5
2014; Mishra et al., 2019; Muley et al., 2015). Induction heating has been mm and a diameter of 20 mm and 23 mm respectively. Enter-
previously used in a few pyrolysis studies (Gauthier et al., 2013; Latifi ing gases are forced to pass through the small spaces (2 mm)
et al., 2014; Muley et al., 2015; Tsai et al., 2009), mainly focused on between the reactor tube and the exchanger walls, for a total
the maximization of the oil fraction, which generally induces tempera- of 3 passes (Fig. 2).
tures below 600 ◦ C. Therefore, the literature regarding pyro-gasification The crucible consists in a metallic cylinder (of 27 mm in
studies using an induction reactor is very scarce. diameter and 55 cm in height), with the bottom part made of
The present work focuses on the design and thermal characteri-
wire mesh, which enables the gas to flow through the sample
zation of a new induction-heated reactor (IHR), for the study of the
particles. It can contain up to 10 g of sample, depending on
thermochemical conversion of solid waste. The characteristics and
the bulk density of the feedstock. A ceramic ring of 5 mm in
thermal behavior of the developed reactor are presented in detail. Pre-
liminary pyrolysis experiments were carried out with solid recovered height acts as a support and places the crucible at the middle
fuel (SRF) samples to validate the setup, and their results are also pre- of the reactor tube, just on top of the preheater. A metallic
sented and compared with tests made in a fluidized bed reactor (FBR). mesh cone of 35 mm height placed below the crucible acts as
Pyrolysis is here studied as the first step of gasification process, and the a diffuser (Fig. 2). Above the sample, an empty region of 320
distribution and characterization of gas and tar products are the main mm high allows the gas phase reactions to take place during
interests. a few seconds.
The reactor is heated externally by an induction setup.
2. Description of the experimental setup The reactor tube is surrounded by a water-cooled copper coil
and procedure inductor of 24 turns disposed in a parallel layout, for a total
height of 420 mm. A generator (HFP 12, EFD induction Gmbh,
2.1. Design requirements 12 kW) supplies energy to the induction circuit. Maximal out-
put value is limited to 75% by an internal setting. Setpoint
The device developed in this work was designed to reach temperature is adjusted with a GEFRAN 2500 PID controller,
reaction thermal conditions close to those of a fluidized bed which is connected to a two-color optical pyrometer (Impac
reactor, frequently used in biomass and waste gasification. IGAR 6, range between 100 and 2000 ◦ C, response time 2 ms).
 Fig. 1 – Experimental setup of the induction heated reactor (IHR).

Fig. 2 – a) Schematic of the reactor tube outlining the wall thermal measuring elements position. b) Cross sectional
illustration of the preheater and sample region, detailing the position of the sample thermocouples in the T shaped rod and
the N2 flow direction.
A quartz tube is placed to create a 5 mm gap between the coil
and the reactor (5 mm gap).

2.2.1. Product collection and analysis modules

Product gas flows upwards and leaves the reactor through the
outlet line, which goes to the condensable trapping system.
The top portion and the outlet line are heated and insulated to
keep temperature above 250 ◦ C, to prevent tar and water con-
densation before reaching the traps. Five gas washing bottles
filled with 2-propanol are used to collect condensable species
(water and tar) present in the stream. Glass wool and glass
beads are used to improve the contact between the gas and
the solvent. The first two bottles are installed in an ice bath
at 0 ◦ C, while the other three are immersed in a carbonic ice
and 2-propanol mixture at −70 ◦ C. One empty bottle is placed Fig. 3 – Temperature evolution at different locations along
at the end of the sampling train. Tar species are subsequently the reactor outer wall.
analyzed and quantified using a gas chromatograph with a
flame ionization detector (GC-MS/FID, Agilent 7890A).
3.1. Dynamic temperature profiles
Non-condensable gases flow towards an online NDIR sen-
The evolution of temperatures at the external reactor tube
sor (Non-Dispersive Infrared Detector), which record the CO,
surface is shown as a function of time in Fig. 3. Tw1 is the
CO2 and CH4 concentrations every 1 s. Total volumetric
closest thermocouple to gas entrance and is placed outside
flowrate of the produced gases is determined by a diaphragm
the heated area, so its temperature was far below the setpoint.
gas meter placed at the outlet of the NDIR analyzer. Simulta-
The wall temperature at the sample height (Tw3) reached the
neously, gases are collected in the Tedlar bag during the whole
setpoint value in 15 s approximately, following a linear behav-
duration of the test (20 min). The test is finished when no
ior with a steep slope of 80 ◦ C/s. At the preheater position
gas concentration change is observed in NDIR. The gas from
(Tw2), the thermal response was slower due to the inertia of
the bag is then accurately analyzed in the ␮GC. An Agilent
the internal elements and to the colder gas flowing along the
3000A chromatograph equipped with four columns is used.
exchanger walls. Temperatures along the empty zone above
The species that can be quantified are O2 , CO2 , CO, CH4 , N2 ,
the sample (Tw4, Tw5) were similar, reaching a peak temper-
H2 , C2 H2 , C2 H4 , C2 H6 , C3 H8 , C6 H6 (benzene), C7 H8 (toluene),
ature just above the setpoint after 60 s followed by a slight
C8 H10 (xylene and ethylbenzene), H2 S, COS, and H2 O in vapor
decrease before stabilizing around 750 ◦ C.
phase. Remaining solid products in the sample crucible are
The thermal profile inside the reactor was established for
weighed and collected for further analyses once the reactor
two cases and is shown in Fig. 4. In the first case, the crucible
has cooled down to ambient temperature.
was empty, while in the second case it was filled with a bed
of chemically inert material (alumina) with an average par-
3. Reactor thermal characterization ticle size of 4 mm. The loading height was about 30 mm. A
significant temperature difference was observed between the
It is common to assume that reaction temperature is that of reactor wall (pyrometer) and the T shaped rod thermocouples
the heating source, or that the temperature inside the reactor for both cases.
is uniform, although this can lead to major errors (Lédé, 2013). Differences between the internal thermocouples (TS1 to
When using induction heating, the the position of the coil TS5) measurements were not significant for the empty case
and of working frequency are crucial to attain a desired tem- (Fig. 4a), which suggests a uniform temperature profile in both
perature profile for a specific workpiece (Hadad et al., 2016). axial and radial directions. For the second case, the fastest
For this setup, the space between the coils, uniform at first, response was recorded above the sample (TS5, z = 235 mm),
was modified to concentrate the magnetic field density along where the thermocouple was subjected to radiant heat inci-
the preheater and the sample zone, where the required power dent from the reactor wall and from the top particles of the
density was higher. The positions of the coil and that of the bed. This explains why the temperature exceeded the setpoint
pyrometer spot were also varied, and the configuration leading value before reaching a stable value. At the middle of the bed
to the shorter response time was kept. (z = 210 mm) the temperature nearest to the hot reactor wall
Heating profiles inside and at the external surface of the (TS4) increased faster than the one at the center (TS2), which
reactor were monitored in various test runs at atmospheric was expected due to the thermal resistance of the bed mate-
pressure, without and in presence of an inert sample. The rial. The gradient in the radial direction was very low when
reactor was heated from ambient temperature to 800 ◦ C compared with the thermal gradient in the axial direction. A
(pyrometer temperature), and then held at stable conditions similar result was observed in previous works of Fernández
for several minutes to ensure that steady state was reached. et al. (2016) and Chatterjee et al. (2017) where negligible radial
Carrier gas (N2 ) flowrate was set at 1 NL/min. Temperatures at temperature change was observed on two radio-frequency
the wall external surface were measured by five K type ther- heated reactors with similar length to diameter ratios. In con-
mocouples (TW1 to TW5) of 0.5 mm in diameter, distributed on trast, temperature differences in the axial direction reached
the outer surface at z = 10, 155, 207, 268, 375 mm as shown in several hundreds of ◦ C. The slowest response was measured at
Fig. 2. F AT shaped rod with five K type thermocouples (TS1 to the bottom of the crucible (TS1). In this point, carrier gas com-
TS5) at different radial (r = 0, 6, 13 mm) and axial positions ing from the preheater meets the colder solid particles and
(z = 185, 210, 235 mm) was introduced into the crucible as then must flow through the empty spaces of the bed, which
displayed in Fig. 2b. implies an additional thermal resistance.
 Fig. 4 – Temperature evolution a) in the empty crucible and b) in the crucible filled with ceramic material.

Table 1 – Chemical composition of the feedstock.

SRF

Moisture (ar) 5.17%

Ash (db) 16.38%
C (db) 48%
H (db) 6%
O (dba ) 26.68%
N (db) 1.33%
S (db) 0.47%
Cl (db) 1.14%

ar: as received; db: dry basis.

a
By difference.

behavior is the opposite for alumina. At 800 ◦ C, ␣ is equal to

180,8 mm2 /s for nitrogen gas and 2.22 mm2 /s for alumina,
which is consistent with the responses observed for the two
case studies in Fig. 4.
Thermal properties for a solid fuel will depend on its com-
position, and its heating behavior will be influenced the by the
changes it is subjected to during its pyrolysis. In the case of lig-
nocellulosic materials like wood, some part of the feedstock
will be released as volatile matter and the rest will remain as
char. Redko et al. (2020) found that the effective diffusivity at
600 ◦ C of raw wood (0.03 mm2 /s) was significantly lower com-
pared to that of its char (0.5 mm2 /s), and that char diffusivity
tended to increase with temperature. For other waste materi-
als like plastics, devolatilization occurs between 400 and 500
◦ C in a single step without char formation, with the excep-

tion of PVC (Ranzi et al., 2016). Gases produced from plastic

pyrolysis present diffusivities in the range of 8–70 mm2 /s at
Fig. 5 – Temperature distribution along the reactor wall at
high temperatures depending on the monomer (Honus et al.,
steady state for a setpoint of 800 ◦ C.
2018). The temperature evolution of a solid waste sample will
be intermediate between the two cases, first it will behave as
For the conditions of this test, Reynolds number was equal the solid, and then once its devolatilization is complete, it will
to 11.75, which indicates a laminar flow regime inside the reac- behave as in the case of gas.
tor. The main heat transfer mechanisms involved during the
heating of the sample were radiation between the inner walls, 3.2. Steady state conditions
and bed particles, followed by a smaller contribution of con-
vection, due to the low gas flowrate. It took about 300 s for the A second two-color optical pyrometer (Impac IGAR 12-LO,
thermocouples of the inner bed to reach their endpoint value, 350−1300 ◦ C range) placed on a moving support was used to
compared to only 100 s for the empty crucible case. measure the temperature profile along the reactor external
The time needed to reach the setpoint temperature surface in steady state conditions. Both runs (with and with-
depends on the thermal properties of the elements inside the out ceramic particle bed) showed the same result, so the mean
reactor. Thermal diffusivity (␣ = k/␳ Cp )] is an indicator of how temperature distribution for the two runs is shown in Fig. 5.
fast a material respond to a temperature change. Thermal dif- The maximum temperature reached inside the sample
fusivity increases with temperature for nitrogen gas, while the bulk (located between 180 and 210 mm) is 808 ◦ C. Tempera-
ture discrepancies were under 10 ◦ C in the bed (TS1 to TS4), 800 ◦ C) is between 2 and 4 s. These conditions were selected
which shows that the temperature is uniform and close to for the pyrolysis test presented in the next section.
the setpoint temperature once the steady state is reached.
The highest wall temperature (TW2 = 825 ◦ C) was measured 4. Pyrolysis tests
at 155 mm, where the preheater device is placed. Above the
sample, an isothermal zone of about 200 mm long was main-
Preliminary pyrolysis experiments were carried out with solid
tained at 750 ◦ C. This empty zone is analog to the freeboard
recovered fuel (SRF) samples to validate the whole setup oper-
zone in a fluidized bed reactor. The temperature progressively
ation and experimental procedure. Their results are presented
decreases above this point since the current density is much
in the present section and compared with tests made in a pilot
lower outside the coil.
scale bubbling fluidized bed reactor (FBR).

4.1. Feedstock
3.3. Reactor modeling

The SRF used for the pyrolysis experiments, was produced

Thermocouple reading inside the tubular reactor can be
from industrial and household waste. The fuel, in the form
affected by the radiative heat transfer between the thermo-
of pellets of 1 cm in height and 5 mm in diameter, was charac-
couple and the surrounding walls. The thermocouple thus
terized by ultimate and proximate analyses. Moisture content
sees surrounding surfaces at a higher or lower temperature
was determined at 105 ◦ C while ash content was measured at
than the real gas temperature. To complete the description
815 ◦ C according to the ISO 1171standard. Elemental composi-
of the gas temperature inside the reactor, a finite element
tion (C, H, N, S) was determined with an ELEMENTAR analyzer.
thermal modelling was performed at steady state. The com-
Feedstock composition is listed in Table 1.
mercial software ANSYS FLUENT R19.3 was used to perform
these calculations.
Reactor geometry was discretized using a 2D- 4.2. Test procedure
axysymmetric approach, including the exchanger and the
sample holder. The model assumed that gas phase consisted The crucible was filled with 2.3 g of dried SRF pellets (24 h at 105
◦ C) and placed in the reactor. Gas washing bottles were placed
solely of nitrogen, and a carrier gas flow rate was set as inlet
boundary condition. Various carrier gas flowrates (0.5 and 4 in the cooling baths and then connected to the reactor outlet.
L/min) were tested to see the preheater performance and the Nitrogen gas (0.5 L/min) was used to flush the entire system
influence of the gas velocity in the temperature profile of the before the test and to carry the produced volatiles. Reactor
reactor. The solid bed of particles was considered by using the was heated up to 800 ◦ C and then held at this temperature
porous media zone condition, which was applied to a specific for about 20 min. All the permanent gases were collected in
flow domain created for the zone inside the crucible. a Tedlar bag, during the whole test and for an additional 20
Properties of gas and solid phases were calculated using min to ensure all the released products were collected. The
temperature dependent functions. Atmospheric pressure was gas meter reading and the composition of permanent gases as
defined as the outlet condition. The experimental profile determined using micro GC were used to calculate the molar
measured previously at the tube surface was implemented amounts of permanent gases. Liquid in the washing bottles
to describe the stationary wall temperature. Reactor walls was sampled and then analyzed by GC-MS/FID. Sample holder
were modeled using no-slip boundary condition for the gas was weighed before and after the test to determine the char
phase. Based on the assumptions listed above, governing yield. The carbon conversion was defined as the ratio between
equations (mass, energy, and momentum) were solved numer- the mass fraction of carbon in each product and the carbon
ically by a finite volume method. The PRESTO (PREssure content in the feedstock. Two tests were performed under the
STaggering Option) was used as the pressure interpolation same conditions and the average values are presented. The
scheme, recommended for problems involving flow through values did not differ by more than 2%.
porous media. The discreet ordinates radiation model was
used to model the radiation heat transfer. The convergence 4.3. Product yields
was verified by monitoring residuals and the mass and energy
balances. The carbon distribution into each of the pyrolysis main prod-
Fig. 6 shows the temperature distribution of the gas phase ucts (permanent gases, tar, char) was determined and is
inside the reactor at steady state conditions for increasing gas presented in Fig. 7. The tar fraction is defined here as all con-
flowrates. At the lowest values, inlet gas was properly heated densable organic compounds with a higher molecular weight
by the heat exchanger, so its temperature remained above 790 than toluene. The results of the present study (named here
◦ C in the crucible zone with a rather homogeneous profile in IHR for “Induction Heating Reactor”) were compared with the
both axial and radial directions. As the N2 flowrates increased, ones obtained in the work of Valin et al. (2019) in which pyrol-
more pronounced temperature gradients were observed in the ysis tests of the same batch of SRF were conducted at 800 ◦ C
crucible zone, especially in the radial direction. Because of in a pilot scale bubbling fluidized bed reactor (FBR) filled with
enhanced convection, the length of the isothermal empty zone olivine as bed material.
above the crucible increased and the maximum temperature In both cases, most of the initial carbon was converted to
position shifted in the direction of the N2 flowrate, as observed gas products. A higher fraction of carbon remained in the char
in the work of Chatterjee et al. (2017). after pyrolysis in the IHR reactor compared with the FBR. This
N2 flowrates of 0.5 and 1 L/min are best suited to achieve is consistent with the results obtained by Efika et al. (2015),
a homogeneous temperature profile in the sample as desired. where RDF samples were pyrolyzed at different heating rates
Under these conditions, the estimated residence time of the in a horizontal tubular reactor. Indeed, higher heating rates
gas in the hot empty zone (temperatures between 750 ◦ C and are encountered in fluidized beds thanks to contact between
Fig. 6 – Contour plots of temperature distribution in the gas phase for different gas flowrates (left). Close up at the
exchanger and sample zone (right).

Fig. 8 – Yield of non-condensable gases from RDF pyrolysis

Fig. 7 – Carbon distribution into products obtained from
at 800 ◦ C for the developed device and a pilot fluidized bed
RDF pyrolysis at 800 ◦ C in the induction heating reactor and
reactor.
a pilot fluidized bed reactor.

hot bed material particles, gas, and feedstock particles, which included H2 , CO2 , CO, CH4 , light hydrocarbons from two to four
favors the carbon conversion to volatiles. In addition to the carbon atoms (C2 H2 , C2 H4 , C2 H6 , C3 H8, C3 H6, C4 H6 -butadiene)
heating rate, the higher temperature and residence time also and light aromatic species like C6 H6 (benzene) and C7 H8
contribute to the cracking of the primary vapors into light (toluene), grouped and named here BTX.
molecular weight hydrocarbons, leading to increased gas yield As stated before, total gas production was higher in the
and lower tar yield. Carbon mass balance closure reached FBR reactor. For the IHR reactor, CO was the most abundant
about 85 wt% in the IHR reactor and was slightly lower for the gas (29.4%), followed by methane (22%) and hydrogen (21.5%).
FBR (78%). The missing carbon fraction could be attributed to The fluidized bed gas presented higher concentrations of H2
condensable products that are deposited in the reactor outlet (31.9%) and CO2 (21.7%) compared to the IHR gas. This could
despite of tracing, and to hydrocarbons that cannot be identi- be due to higher residence time as well as the presence of
fied or quantified in the GC-MS/FID analysis. bed material in the FBR, which favored hydrocarbon break-
down reactions resulting in more H2 . Olivine bed improves
4.4. Gas products the reforming of phenol and other tars, as seen in the work of
Nitsch et al. (2013). The contents of methane, light hydrocar-
Produced gas yield and its volumetric composition are com- bons and BTX were slightly lower in the FBR. A similar trend
pared in Fig. 8 for both setups. In this study, gas species was observed by Pasel and Wanzl (2003), who compared the
 performed with raw materials used as model constitutents
of wastes (such as plastic polymers, woody biomass, paper
or cardboard), and their mixtures. Results obtained in the
IHR reactor, carried out under controlled and well character-
ized conditions, will then be useful for the development and
improvement of prediction models of the reaction products,
and thus contribute to the understanding of pyrolysis and
gasification at various scales.
Fig. 9 – Composition of tar produced from the pyrolysis of
SRF in the developed device and a pilot fluidized bed Declaration of competing interest
reactor.
The authors declare that they have no known competing com-
benzene yield from shredded waste pyrolysis in various types
mercial interests or personal relationships that could have
of equipment, including a tubular and a fluidizing reactor.
appeared to influence the work reported in this paper.

4.5. Tar products

References
The condensable fraction recovered in the gas washing bottles
contains numerous organic species, which were then analyzed Aluri, S., Syed, A., Flick, D.W., Muzzy, J.D., Sievers, C., Agrawal,
P.K., 2018. Pyrolysis and gasification studies of model refuse
and quantified by GC-MS/FID. Here, the main tar species from
derived fuel (RDF) using thermogravimetric analysis. Fuel
pyrolysis in both reactors are shown in Fig. 9. Composition of Process. Technol. 179 (Oct), 154–166,
tar produced from the pyrolysis of SRF in the developed device http://dx.doi.org/10.1016/j.fuproc.2018.06.010.
and a pilot fluidized bed reactor. Fig. 9, and are classified in Barr, M.R., Volpe, R., Kandiyoti, R., 2019. Influence of reactor
four groups based on the number of aromatic rings in their design on product distributions from biomass pyrolysis. ACS
structure. Sustain. Chem. Eng. 7 (Aug (16)), 13734–13745,
http://dx.doi.org/10.1021/acssuschemeng.9b01368.
Longer residence times and the presence of bed material
Benedikt, F., Schmid, J.C., Fuchs, J., Mauerhofer, A.M., Müller, S.,
possibly favored the extent of secondary reactions in the flu-
Hofbauer, H., 2018. Fuel flexible gasification with an advanced
idized bed reactor, which decreased the total amount of tar 100 kW dual fluidized bed steam gasification pilot plant.
when compared to the IHR. Most of the tar species generated Energy 164 (Dec), 329–343,
in the IHR reactor were in the first group, consisting mostly http://dx.doi.org/10.1016/j.energy.2018.08.146.
of styrene (74%), and followed by indene and phenol. In the Chatterjee, S., Houlding, T.K., Doluda, V.Yu., Molchanov, V.P.,
case of the fluidized bed, naphthalene, with two rings, was Matveeva, V.G., Rebrov, E.V., 2017. Thermal behavior of a
catalytic packed-bed milli-reactor operated under radio
the most abundant component (70%). Heavier polycyclic aro-
frequency heating. Ind. Eng. Chem. Res. 56 (Nov (45)),
matic hydrocarbons (PAHs) such as anthracene and pyrene (3
13273–13280, http://dx.doi.org/10.1021/acs.iecr.7b01723.
and 4 rings) were present in the tar from the FBR, while their Cortazar, M., Lopez, G., Alvarez, J., Arregi, A., Amutio, M., Bilbao,
amount for the IHR case was negligible. High temperatures J., Olazar, M., 2020. Experimental study and modeling of
(above 700 ◦ C) promote the appearance of PAHs, as a product biomass char gasification kinetics in a novel
of aromatization and rearrangement reactions of aliphatic and thermogravimetric flow reactor. Chem. Eng. J. 396 (Sep),
monocyclic compounds (Pasel and Wanzl, 2003). 125200, http://dx.doi.org/10.1016/j.cej.2020.125200.
Daouk, E., Sani, R., Pham Minh, D., Nzihou, A., 2018.
Thermo-conversion of Solid Recovered Fuels under inert and
5. Conclusions oxidative atmospheres: gas composition and chlorine
distribution. Fuel 225 (Aug), 54–61,
An induction-heated reactor was developed to perform pyrol- http://dx.doi.org/10.1016/j.fuel.2018.03.136.
ysis experiments of waste derived fuels and its constituent Efika, E.C., Onwudili, J.A., Williams, P.T., 2015. Products from the
materials. Some of the features of the developed device high temperature pyrolysis of RDF at slow and rapid heating
rates. J. Anal. Appl. Pyrolysis 112 (Mar), 14–22,
included a gas preheating system, high heating rates and the
http://dx.doi.org/10.1016/j.jaap.2015.01.004.
possibility of recovering and analyzing all reaction products. Efika, C.E., Onwudili, J.A., Williams, P.T., 2018. Influence of heating
Experimental and CFD results provided a complete descrip- rates on the products of high-temperature pyrolysis of waste
tion of the temperature profiles for the surface, the sample wood pellets and biomass model compounds. Waste Manag.
and the gas inside the reactor. Heat provided by the induction 76 (Jun), 497–506,
circuit enabled to maintain a uniform temperature inside the http://dx.doi.org/10.1016/j.wasman.2018.03.021.
Fernández, J., Sotenko, M., Derevschikov, V., Lysikov, A., Rebrov,
sample carrier gas flowrates between 0.5 and 1 NL/min.
E.V., 2016. A radiofrequency heated reactor system for
The IHR was used to perform the pyrolysis of a SRF sam-
post-combustion carbon capture. Chem. Eng. Process. Process
ple and the results were compared with tests made in a pilot Intensif. 108 (Oct), 17–26,
scale fluidized bed reactor. Reaction conditions in the IHR reac- http://dx.doi.org/10.1016/j.cep.2016.07.004.
tor were close to intermediate to fast pyrolysis (heating rate Garcés, D., Díaz, E., Sastre, H., Ordóñez, S., González-LaFuente,
75 ◦ C/s, volatiles residence time ≈ 4 s). Char yield indicates J.M., 2016. Evaluation of the potential of different high
that the heating rate in the solid sample is lower in the IHR calorific waste fractions for the preparation of solid recovered
fuels. Waste Manag. 47 (Jan), 164–173,
compared to the FBR.
http://dx.doi.org/10.1016/j.wasman.2015.08.029.
Main permanent gas products were similar for the two
Gauthier, G., Melkior, T., Grateau, M., Thiery, S., Salvador, S., 2013.
reactors; however, residence time and the presence of the bed Pyrolysis of centimetre-scale wood particles: new
material had an influence on the extent of secondary reac- experimental developments and results. J. Anal. Appl.
tions, which modified the distribution of the permanent gas Pyrolysis 104 (Nov), 521–530,
species and tar products. In the future, experiments will be http://dx.doi.org/10.1016/j.jaap.2013.05.017.
Hadad, Y., Kochavi, E., Levy, A., 2016. Inductive heating with a downdraft fixed bed and bubbling fluidized bed reactors.
stepped diameter crucible. Appl. Therm. Eng. 102 (Jun), Waste Biomass Valorization (Aug),
149–157, http://dx.doi.org/10.1007/s12649-018-0431-6.
http://dx.doi.org/10.1016/j.applthermaleng.2016.03.151. Pasel, C., Wanzl, W., 2003. Experimental investigations on reactor
Henkel, C., 2014. A study of induction pyrolysis of lignicellulosic scale-up and optimisation of product quality in pyrolysis of
biomass for the production of bio-oil. LSU Masters Theses shredder waste. Fuel Process. Technol. 80 (Jan (1)), 47–67,
(Jan) [Online]. Available: http://dx.doi.org/10.1016/S0378-3820(02)00187-X.
https://digitalcommons.lsu.edu/gradschool theses/447. Porshnov, D., Ozols, V., Ansone-Bertina, L., Burlakovs, J., Klavins,
Honus, S., Kumagai, S., Fedorko, G., Molnár, V., Yoshioka, T., 2018. M., 2018. Thermal decomposition study of major refuse
Pyrolysis gases produced from individual and mixed PE, PP, PS, derived fuel components. Energy Procedia 147 (Aug), 48–53,
PVC, and PET—part I: production and physical properties. Fuel http://dx.doi.org/10.1016/j.egypro.2018.07.032.
221 (Jun), 346–360, http://dx.doi.org/10.1016/j.fuel.2018.02.074. Ranzi, E., Faravelli, T., Manenti, F., 2016. Pyrolysis, gasification,
Hwang, I.-H., Kobayashi, J., Kawamoto, K., 2014. Characterization and combustion of solid fuels. Advances in Chemical
of products obtained from pyrolysis and steam gasification of Engineering, vol. 49. Elsevier, pp. 1–94.
wood waste, RDF, and RPF. Waste Manag. 34 (Feb (2)), 402–410, Redko, T., Volford, A., Marek, E.J., Scott, S.A., Hayhurst, A.N., 2020.
http://dx.doi.org/10.1016/j.wasman.2013.10.009. Measurement of the times for pyrolysis and the thermal
Latifi, M., Chaouki, J., 2015. A novel induction heating fluidized diffusivity of a pyrolysing particle of wood and also of the
bed reactor: its design and applications in high temperature resulting char. Combust. Flame 212 (Feb), 510–518,
screening tests with solid feedstocks and prediction of http://dx.doi.org/10.1016/j.combustflame.2019.10.024.
defluidization state. AIChE J. 61 (5), 1507–1523, Robinson, T., Bronson, B., Gogolek, P., Mehrani, P., 2016. Sample
http://dx.doi.org/10.1002/aic.14749. preparation for thermo-gravimetric determination and
Latifi, M., Berruti, F., Briens, C., 2014. Non-catalytic and catalytic thermo-gravimetric characterization of refuse derived fuel.
steam reforming of a bio-oil model compound in a novel Waste Manag. 48 (Feb), 265–274,
‘Jiggle Bed’ reactor. Fuel 129 (Aug), 278–291, http://dx.doi.org/10.1016/j.wasman.2015.11.018.
http://dx.doi.org/10.1016/j.fuel.2014.03.053. Saadatkhah, N., Garcia, A.C., Ackermann, S., Leclerc, P., Latifi, M.,
Lédé, J., 2013. Biomass fast pyrolysis reactors: a review of a few Samih, S., Patience, G.S., Chaouki, J., 2020. Experimental
scientific challenges and of related recommended research methods in chemical engineering: thermogravimetric
topics. Oil Gas Sci. Technol. – Rev. D’IFP Energ. Nouv. 68 (Sep analysis—TGA. Can. J. Chem. Eng. 98 (1), 34–43,
(5)), 801–814, http://dx.doi.org/10.2516/ogst/2013108. http://dx.doi.org/10.1002/cjce.23673.
Leion, H., Frick, V., Hildor, F., 2018. Experimental method and Samih, S., Chaouki, J., 2015. Development of a fluidized bed
setup for laboratory fluidized bed reactor testing. Energies 11 thermogravimetric analyzer. AIChE J. 61 (1), 84–89,
(Oct (10)), 2505, http://dx.doi.org/10.3390/en11102505. http://dx.doi.org/10.1002/aic.14637.
Marshall, A.J., Wu, P.F., Mun, S.-H., Lalonde, C., 2014. Commercial Tsai, W.-T., Chang, J.-H., Hsien, K.-J., Chang, Y.-M., 2009.
application of pyrolysis technology in agriculture. Am. Soc. Production of pyrolytic liquids from industrial sewage sludges
Agric. Biol. Eng. Annu. Int. Meet. 2014 ASABE 2014 5, in an induction-heating reactor. Bioresour. Technol. 100 (Jan
3868–3886. (1)), 406–412, http://dx.doi.org/10.1016/j.biortech.2008.06.013.
Mishra, H., Patidar, B., Pante, A.S., Sharma, A., 2019. Valin, S., Ravel, S., Pons de Vincent, P., Thiery, S., Miller, H., 2019.
Mathematical modelling, simulation and experimental Fluidized bed air gasification of solid recovered fuel and
validation of resistance heating and induction heating woody biomass: influence of experimental conditions on
techniques for E-waste treatment. IET Electr. Power Appl. 13 product gas and pollutant release. Fuel 242 (Apr), 664–672,
(4), 487–493, http://dx.doi.org/10.1049/iet-epa.2018.5535. http://dx.doi.org/10.1016/j.fuel.2019.01.094.
Muley, P.D., Henkel, C., Abdollahi, K.K., Boldor, D., 2015. Pyrolysis Zhang, L., Bao, Z., Xia, S., Lu, Q., Walters, K., 2018. Catalytic
and catalytic upgrading of pinewood sawdust using an pyrolysis of biomass and polymer wastes. Catalysts 8 (Dec
induction heating reactor. Energy Fuels 29 (Nov (11)), (12)), 659, http://dx.doi.org/10.3390/catal8120659.
7375–7738, http://dx.doi.org/10.1021/acs.energyfuels.5b01878. Zhou, H., Long, Y., Meng, A., Li, Q., Zhang, Y., 2015.
Nilsson, S., Gómez-Barea, A., Cano, D.F., 2012. Gasification Thermogravimetric characteristics of typical municipal solid
reactivity of char from dried sewage sludge in a fluidized bed. waste fractions during co-pyrolysis. Waste Manag. 38 (Apr),
Fuel 92 (Feb (1)), 346–353, 194–200, http://dx.doi.org/10.1016/j.wasman.2014.09.027.
http://dx.doi.org/10.1016/j.fuel.2011.07.031. Zhu, J., Jin, L., Luo, Y., Hu, H., Xiong, Y., Wei, B., Wang, D., 2020.
Nitsch, X., Commandré, J.-M., Clavel, P., Martin, E., Valette, J., Fast co-pyrolysis of a massive Naomaohu coal and cedar
Volle, G., 2013. Conversion of phenol-based tars over olivine mixture using rapid infrared heating. Energy Convers.
and sand in a biomass gasification atmosphere. Energy Fuels Manage. 205 (Feb), 112442,
27 (Sep (9)), 5459–5465, http://dx.doi.org/10.1021/ef400817z. http://dx.doi.org/10.1016/j.enconman.2019.112442.
Park, S.-W., Lee, J.-S., Yang, W.-S., Alam, M.T., Seo, Y.-C., 2018. A
comparative study of the gasification of solid refuse fuel in