Tar Reforming Under A Microwave Plasma Torch
Tar Reforming Under A Microwave Plasma Torch
Tar Reforming Under A Microwave Plasma Torch
pubs.acs.org/EF
ABSTRACT: Because of the scarcity of nonrenewable natural resources, such as petroleum and natural gas, the use of biofuel is
needed. Gasification is a major process used to obtain renewable fuels from biomass; however, the gas cleaning system is a
constraint for its broad utilization. During the pyrolysis process, a mixture of organic compounds in the gas phase is produced
and must be removed from the gases before it is used in the most practical applications. In order to remove such organic
compounds, which are known as tar, large, sophisticated, problematic, and expensive gas cleaning systems are added to the
gasifier gas exit. Previous papers have shown that the plasma torch has the potential to destroy produced tar, being a simpler and
less-expensive system than traditional gas cleaners. This work presents a qualitative and quantitative evaluation of a microwave
plasma system running on tar destruction and its reforming. In order to evaluate a 1 kW microwave plasma system performance,
an apparatus was developed and installed at ITA Laboratory of Plasmas and Processes (LPP-ITA). The system runs at
atmospheric pressure with nitrogen and argon as carrier gas under a large range of flow rates. Experiments were performed using
a gas mixture of N2, H2O, ethanol, and tar at controlled concentration in order to simulate the gases produced by a gasifier. The
injected tar was obtained from pine pyrolysis and characterized for energy purposes. In order to reduce tar viscosity, it was
diluted in commercial ethanol (92.5% ethanol and 7.5% water) and its concentration varied from 0.8 gtar/Nmgas3 to 4.2 gtar/
Nmgas3. Species formed in the microwave plasma torch were identified using an optical spectrometer. The reactor exit gases had
their composition evaluated on tar content as well as for noncondensable gases. As a result, this paper shows that no tar content
was detected at the reactor outlet, indicating that all supplied tar was destroyed in the plasma reactor. The main detected
products were CO and solid carbon (C(s)). Furthermore, neither NO nor CO2 were detected, and an indication of H2 formation
was obtained. This paper concludes that the microwave plasma system is capable of destroying and reforming tar efficiently and
produces mainly H2, CO, O2, and C(s) as byproducts.
© 2012 American Chemical Society 1174 dx.doi.org/10.1021/ef301399q | Energy Fuels 2013, 27, 1174−1181
Energy & Fuels Article
Figure 1. Apparatus scheme adopted in this work for the tar reforming with a microwave plasma torch.
torch was started up only with argon flowing through it. The After passing through the second tube furnace, the gas
reactor then was supplied with a known mixture of argon, mixture was driven to the microwave plasma reactor. In the
nitrogen, and tar solution (ethanol + water + pine tar). The entrance of the microwave plasma reactor, there is a Type K
reason for the presence of argon upon ignition and afterward thermocouple connected to a digital display, which measures
was to provide stability to the torch (and avoid subsequent doped gas temperature at the reactor entrance.
torch extinction), because of the fact that argon is easier In order to avoid tar condensation and subsequent
ionized, when compared to other gases. The reason for using obstruction in the pipe feeding, tar solution was injected
nitrogen instead of argon was to simulate gasifier exhaust gases, when the gas temperature at the reactor inlet was equal to or
which are mostly comprised of nitrogen. higher than 250 °C. This temperature value was chosen after
2.1. Doping and Evaporation. Figure 1 illustrates the experimental evaluation for tar solution boiling temperature.
experimental apparatus. Nitrogen and argon go through flow The doped gas mixture then enters in the microwave plasma
meters, where they have their flow rates individually registered reactor, where tar breakdown was expected to occur, as well as
and controlled with valves. After passing through flow meters, new chemical species formation.
gases mix themselves and go to a 1/4-in. copper pipe coil 2.2. Microwave Plasma Reactor. The microwave plasma
located inside a tube furnace thermally heated with 3500 W reactor utilized in this work, with a maximum power of 3000 W
electrical resistances. In order to avoid heat losses, this furnace and source voltage of 5500 V, is illustrated in Figure 1. Its
was insulated with rock wool and glass wool covered with magnetron converts the electrical current into microwave. A
aluminum foil. Standard operation procedure keeps the gas circulator is located downstream from the magnetron source, to
temperature at the furnace exit heated to temperatures over 300 prevent the electromagnetic current from returning and damage
°C. After passing through the first tube furnace, argon− the magnetron. The circulator has a system to measure incident
nitrogen mixture receives tar solution from a syringe pump and reflected power, which is intended to evaluate the
(manufactured by KDS Scientific, Model KDS100, with an microwave application efficiency into the gas passing through
error of 1%). This 10-mL syringe pump dopes the carrier gas the reactor. Downstream from the circulator, the microwaves
with tar solution comprised of 63.49% ethanol (mass fraction), are conducted through a waveguide to a quartz tube, 10 mm
which is responsible for tar dilution; 6.51% water (oxygen and ID, where they were absorbed by the gas coming from the
hydrogen donor); and 30% pine tar. feeding system, on swirl way, ionizing a fraction of molecules in
The heated mixture of argon and nitrogen drags and this gas, thereby generating the plasma torch. In order to
evaporates the tar solution injected by the syringe pump while maximize tar breakdown and facilitate plasma generation, the
it flows in the second furnace. The second tube furnace impedance of the plasma region is important to approach the
increases the doped gas temperature up to ∼350 °C. It is impedance from the microwave generator (magnetron). This is
responsible for completing the tar evaporation process, seeking accomplished by using a movable piston and stubs.
to avoid droplet presence in the plasma torch. The second tube The microwave plasma reactor ignition system consists of a
furnace was constructed similarly to the first one. tungsten rod inside the quartz tube in the plasma region
1175 dx.doi.org/10.1021/ef301399q | Energy Fuels 2013, 27, 1174−1181
Energy & Fuels Article
Figure 2. Apparatus scheme for the tar measurement system applied in this work.
(shown in Figure 1). The purpose of such a tungsten rod is to The retained tar quantification was performed as follows:
concentrate the electromagnetic waves and start the plasma before being used, the gas washing flask was individually
torch. Upon ignition, maximum power is imposed to the flow weighed with 100 mL of isopropanol. After using these bottles,
(3000 W) during a short period of time (on the order of the mixture of tar and isopropanol contained in each individual
milliseconds). After this period, the system reduces its power to bottle was weighed again, and the difference in flask weight
the chosen prerequested level and holds it constant and steady before and after experiment is the amount of tar mass trapped.
(1000 W). The power can be modified later, without It is important to keep in mind that the reactor eluded gas has
interruptions, during the microwave plasma reactor operation. no water in its composition.
2.3. Tar Measurement System. Downstream from the The borosilicate filter was dried in a drying oven for 1 h
microwave plasma reactor is the tar concentration measure- (time provided by the manufacturer) before being used in the
ment system. The applied tar measurement system, shown in tar measurement system. After passing through the laboratory
Figure 2, with a lower detection limit of 0.01gtar/mgas3, was drying oven, the borosilicate filters were weighed. After the
based on the procedure9,10 and successfully utilized.11 experiment, the borosilicate filter was weighed again and the
The exhaust gases from the plasma torch were fed into the difference between these weights was added to the tar mass
trapped in the washing flasks, thereby providing the total mass
tar filtration system at a specific mass flow rate. Such flow was
of retained tar.
controlled through a valve, and flow surplus was sent to the
The volume shown in the gas meter after the test is
alternative output. The gas sample then flows through a tar subtracted from the amount shown before the test was
borosilicate filter (1 μm mesh filter with efficiency above 99%). performed. The difference between such volumes is the gas
Its external wall is electrically heated at 200 °C, using a 120-W volume analyzed by the tar measurement system. Using the
electrical-resistance-type collar, thereby preventing tar from values obtained from the manometer and the thermocouple at
condensing on the filter holder walls. In this filter, mainly the gas meter, the volume of the analyzed gas was converted
primary tars were trapped and removed from the hot gas. from mgas3 to Nmgas3. The total amount of retained tar in the tar
The next step was trapping the remaining tar content measurement system was divided by the gas volume (gtar/
through one or up to six gas washing flasks (Schout, 500 mL Nmgas3), as shown in eq 1.
with 100 mL of isopropanol in each flask). The quantity of gas
m
washing flasks depends upon the tar concentration in the gas. C tar = tar
The flasks with isopropanol retain the remaining tars that will Vgas (1)
be quantified through weight measuring. The last flask of 500
mL contains silica gel and works as tar presence warning, 2.4. Optical Spectroscopy and Tar Characterization.
During the species ionization process present in the plasma
indicating need to add new washing flasks and avoid tar
torch, these species have their electrons excited. The excited
presence in the vacuum pump. The gas washing flasks were
electrons return to the standard state, thereby emitting
kept at temperatures of approximately −5 °C with an 18-L photons. The color and tone of the photons are characteristics
ethylene glycol cooler system, cooling the gas and facilitating of each chemical species of the plasma torch present, allowing
tar condensation. species identification through optical emission spectroscopy.
The cooled gas, which is now free from tar, passes through a The optical spectroscopy utilized in this experiment was an
vacuum pump (with a maximum flow of 120 L/min). After optical emission spectrometer (Model UV−vis, Ocean Optics
passing through such vacuum pump, the gas flows through a USB4000) with a resolution of 1.5 nm and operating range in
meter with a Type K thermocouple connected to a temperature the ultraviolet (UV) and visible (Vis). The spectrum was
display and a manometer (pressure gauge minihelic II model obtained by an OOIBase32 program.
with a measurement range of pressure between 0 and 100 mm Chromatographic analysis was performed in the sample tar
of H2O). Subsequently, the gas without tar was driven to a gas solution aiming to identify organic species in the tar. The
analyzer (greenline GL8000) with detection ranges from 20 chromatograph used was a Model GCT-Premier system
ppm up to 8000 ppm for CO and from 250 ppm up to 250 000 manufactured by Waters, and its main parameters are given
ppm for O2) or released to the atmosphere. as follows: column, HP-5 (5% phenylmethylsiloxane) MS 30 m
1176 dx.doi.org/10.1021/ef301399q | Energy Fuels 2013, 27, 1174−1181
Energy & Fuels Article
of long × 0.25 mm ID; initial temperature, 40 °C; column Table 1. Main Operational Parameters Adopted during the
temperature, 280 °C; injector temperature, 270 °C; detector Experiments
temperature, 280 °C; helium flow, 1 mL/min; constant
tar solution (mL/h) tar concentration (gtar/Nmgas3)
pressure, 80 kPa; and library peak identification used, NIST
(National Institutes for Standards and Technology). 6 1.7
In order to qualify the atomic composition present in the 9 2.5
used tar, an ultimate analysis was conducted. The apparatus 12 3.3
used for such analysis was an Ultimate Analyzer 2400 Series II 15 4.2
CHNS/O system and an analytical balance of six decimal
figures (Model AD-6), both manufactured by Perkin−Elmer.
of the plasma reactor was ∼250 °C. Nitrogen and argon flow in all
The EBMA laboratory also provided measurement for used the experiments were kept constant at 8.8 and 3.8 L/min, respectively.
gross heating value (GHV) of the used tar. The apparatus used Throughout the operation, the power reflected and not absorbed by
for such analysis was a calorimeter manufactured by Ika Werke the plasma torch did not exceed 0.5% of the power supplied by the
(Model C 2000). microwave plasma reactor.
Figure 3. Result from chromatographic gaseous mass spectrometry analysis (CG/MS) of the tar solution used in this experiment.
rate, the quartz tube wall saturates so fast that it quenches the
plasma. The beginning of instabilization is shown in Figure 7
(CO curve after tar solution of 12 mL/h).
The presence of hydrocarbons (CxHy) was not detected in
the exhaust gases, thereby reinforcing the hypothesis that tar,
after being reformed in a plasma torch microwave, results in
small chain species (diatomic or triatomic). The presence of
SO2 concentration (max average concentration of 212.48 ppm)
was measured, and the presence of either NO or CO2 was not
verified.
4.3. Inlet−Exit Species Mass Balance. In order to
validate the data obtained, an investigation on the absence of
NO and CO2, and possible formation of H2 molecules, a mass
balance was conducted. Table 3 shows the inlet mass flow rates
for all listed species, and Table 4 contains the exit mass flow
rates only for the species listed. Here, only the tar solution
atom population was considered, after argon and nitrogen were
determined to play an inert role. In Table 4, the solid carbon
flow rate was obtained by the difference between the carbon
flow rate from tar solution and that from the eluded CO
measured. In the same table, the adsorbed oxygen flow rate was
obtained using the difference between the oxygen flow rates
from tar solution and that from the eluded CO, SO2, and O2
measured.
As written above, the approach adopted in this work was to
Figure 5. Results from optical emission spectroscopy for inlet gas with keep the mass ratio between atoms constant and independent
tar. of the tar solution flow rate. This can be inferred from the last
five columns in Table 3. In Table 4, the last three columns on
the right show the atoms mass flow rate evaluated only with the
measured species downstream from the reactor (CO, SO2, O2,
and C(s)). One can see that an increase in tar solution flow
increases the CO formation, but this growth is not linear,
indicating excess of free carbon available for reaction and
having, as a consequence, an increased solid carbon formation.
The relationships between C and O for each tar solution flow
gap in Table 3 show an inlet C/O ratio that is constant and
independent of the tar solution flow rate applied (C/O = 1.3).
Therefore, the reactant residence time in the plasma region
influences the formation of CO and O2. An increase in the tar
solution flow rate implies a reduction of reactant residence time
in the plasma reactor, decreasing the O2 formation and
increasing the CO formation. Another observation that occurs
with the increase of the tar mixture flow, thereby reducing the
residence time in the reactor, is SO2 formation. It seems to go
Figure 6. Results from optical spectrum formed after the destruction through a maximum of 9 mL/h.
of water molecules (only). Observing the oxygen atom flow adsorbed by carbon
contained in the quartz tube, which is shown in Table 4 (O-
came from ethanol, water, and tar. Furthermore, the approach adsorbed), there is a very low flow, which is null at low tar
adopted in this work keeps the inlet C/O and C/H ratios solution flows, meaning that all the oxygen present in the tar
constant and independent of the tar solution flow rate applied solution reacts with carbon and the remainder, with sulfur and
(C/O = 1.3 and C/H = 4.6). Therefore, an increase in the tar oxygen, forming SO2 and O2, respectively. This fact can explain
solution flow rate implies only a reduction in reactant residence the absence of both CO2 and NO downstream from the
time in the plasma reactor. reactor. Another important fact is that little oxygen is available
The initially expected result for Figure 7, once tar mixture to react with atomic hydrogen in order to form H2O molecules.
rate increases, is an increase on CO and O2 mass fraction as Moreover, as shown in this paper, there was no unburned
well; but this was not the result that was obtained. Possibly, it is hydrocarbon presence downstream from the reactor; therefore,
consequence of the reactant residence time: Long residence hydrogen atoms do not react with available carbon. Therefore,
time favors O2 formation and low residence time favors CO the hydrogen atoms may have reacted with each other in order
formation. to form hydrogen molecules (H2), instead of reacting with
This experiment was conducted with lack of oxygen to either oxygen or carbon atoms. Because of a lack of availability
produce CO (C/O = 1.68, after ultimate analysis); of equipment in the laboratory during this work, hydrogen gas
consequently, free carbon turned solid and was deposited on was not quantified. The main products measured were C(s), O2,
the quartz tube. In addition, above a certain tar solution flow CO, SO2, and possibly H2. With the exception of SO2
1179 dx.doi.org/10.1021/ef301399q | Energy Fuels 2013, 27, 1174−1181
Energy & Fuels Article
Figure 7. Concentrations of CO and O2 (mole fraction) measured downstream from the microwave plasma reactor, relative to the inlet tar flow rate.
ḿ (g/min) 9 22.319 14.5 7.669 0.0079 0.0971 0.0450 0.07249 0.05567 0.00301 0.00308 0.01575
Y (%) 9 100 64.97 34.35 0.0437 0.4266 0.2016 0.07249 0.05567 0.00301 0.00308 0.01575
ḿ (g/min) 12 22.369 14.5 7.669 0.0105 0.1295 0.0600 0.09666 0.07422 0.00401 0.00411 0.02100
Y (%) 12 100 64.83 34.28 0.0582 0.5676 0.2682 0.09666 0.07422 0.00401 0.00411 0.02100
ḿ (g/min) 15 22.419 14.5 7.669 0.0131 0.1618 0.0750 0.12082 0.09278 0.00501 0.00514 0.02625
Y (%) 15 100 64.68 34.20 0.0726 0.7079 0.3345 0.12082 0.09278 0.00501 0.00514 0.02625
ḿ (g/min) 9 0.13719 0.06634 0.00938 0.01750 0.04397 0.00000 0.02853 0.06000 0.00469
X (ppm) 9 3435.50 212.48 792.96
ḿ (g/min) 12 0.17455 0.07530 0.00733 0.01562 0.06428 0.01201 0.03238 0.06221 0.00367
X (ppm) 12 3899.50 166.13 707.83
ḿ (g/min) 15 0.21678 0.07752 0.00636 0.01473 0.08749 0.03068 0.03333 0.06210 0.00318
X (ppm) 15 4014.50 144.00 667.65
molecules, these products are the same that are obtained when from the plasma reactor, but not immediately downstream. The
destroying and reforming carbon particles through the use of a reactor exit was coupled to the tar sampling line, but
microwave plasma torch.15,16 occasionally the reactor exit gas temperature was obtained
4.4. Energy Consumption. During the experiment, the gas (∼340 °C) when the reactor-sampling line coupling was
temperature was regularly measured immediately upstream disassembled. Without a regular measurement of gas temper-
1180 dx.doi.org/10.1021/ef301399q | Energy Fuels 2013, 27, 1174−1181
Energy & Fuels
■
Article
ature immediately after the plasma reactor, the energy balance ACKNOWLEDGMENTS
was not performed. Instead, the specific energy consumption Authors wish to thanks the Aeronautical Institute of
was evaluated through the ratio between the inlet electric Technology (ITA), the Federal University of Para (UFPA),
power and the mass of tar destroyed. Despite experimental the University of Campinas (UNICAMP), Federal Agency for
observations indicating that tar was destroyed through a Support and Evaluation of Graduated Education (CAPES-
microwave effect, further experiments must validate such a Procad NF), the Polaris, and the National Council for Scientific
conclusion against a thermal effect. and Technological Development (CNPq-Project No. 578106/
The specific energy consumption for a plasma torch can be 2008-6) for allowed grants and support given, as well as
defined as the ratio between incident power and the destroyed doctorate and master fellowships for R.M.E. and B.A.P.C.
■
tar mass flow rate (J/kg). Under the best operating conditions
during all experiments in this work, 15 mL/h of tar solution REFERENCES
(4.5 g/h of tar) was destroyed with 995 W of microwave power (1) Jasiński, M.; Dors, M.; Mizeraczyk, J. Application of Atmospheric
consumed, meaning 796 kJ/gtar. This work did not try to reduce Pressure Microwave Plasma Source for Production of Hydrogen via
power; therefore, it is possible to obtain lower specific energy Methane Reforming. Eur. Phys. J., D 2008, 170−183.
consumption. (2) Nair, S. A.; Pemen, A. J. M.; Yan, K.; Van Heesh, E. J. M.;
Pitazinski, K. J. Drinkenburg, A. A. H. A High Temperature Pulsed
Corona Plasma System for Tar Removal From Biomass Derived Fuel
5. CONCLUSIONS
Gas. J. Electrostat. 2004, 61, 117−127.
Considering the results obtained with Greenline GL8000 (the (3) Gower, S. A. Development of a High Power Microwave Plasma
absence of unburned hydrocarbons) and keeping in mind that Beam Applicator. Sci. Instrum. 2006, 72 (11), 604−611.
the tar measurement system has a lower limit of detection for (4) Stonies, R.; Schermer, S.; Voger, E.; Broekaert, J. A. C. A New
tar concentration of 0.01 gtar/Nmgas3 and a tar concentration up Small Microwave Plasma Torch. Plasma Sources Sci. Technol. 2004,
to 4.2 gtar/Nmgas3 was imposed with no tar concentration 604−611.
(5) Li, C.; Suzuki, K. Tar Property, Analysis, Reforming Mechanism and
detected in the gas downstream from the reactor, the Model for Biomass Gasification Review; Ecotopia Science Institute:
conclusion is that a microwave plasma torch operating at Nagoya, Japan, 2007; pp 179−173.
ambient pressure can break all tar molecules into smaller (6) Fourcault, A.; Marias F.; Michon U. Modelling of Thermal
species. Removal of Tars in a High Temperature Stage Fed by a Plasma Torch.
Results from the optical spectrum, as well as those from a gas Biomass Bioenergy 2010, 34, 1363−1374.
analyzer and calculated in mass balance, drive one to the (7) Jess, A. Mechanism and Kinetics of Thermal Reactions of
conclusion that the destroyed tar was converted mainly to CO, Aromatics Hydrocarbons from Pyrolysis of Solid Fuels. Fuel 1996,
O2, solid carbon (C(s)), and probably into H2 formation. 1441−1448.
Hydrogen gas formation was inferred by comparison between (8) Uhm, H. S.; Kim, J. H.; Hong, Y. C. Disintegration of Water
Molecules in a Steam-plasma Torch Powered by Microwaves. Phys.
the results provided from the gas analyzer and the mass
Plasmas: Korea 2007, 14.
balance, along with the results shown by optical spectroscopy. (9) Knoef, H.; Verenum, G.; Good, J. et al. Sampling and Analysis of
Reactant resident time in the plasma region influences the Tar and Particles in Biomass Producer Gases; Technical Report Casella
product formation: low residence time favors CO formation Cre energy CEN BT/TF 143, Switzerland, 2005; 44 pp.
and high residence time favors O2. The results shown in this (10) Simell, P.; Stahlberg, P.; Kurkelaa, E.; Albrechtb, J.; Deutsch, S.;
paper can validate the proposed reaction pathways for the tar Sjöströmd, K. Provisional Protocol for the Sampling and Analysis of
conversion into proposed smaller species,6 i.e., the microwave Tar and Particulates in the Gas from Large-scale Biomass Gasifiers,
plasma source, at atmospheric pressure, is capable of reducing Version 1998. Biomass Bioenergy 1999, 19−38.
tar to smaller species, such as CO, H2, and C(s). (11) Hla, S. S. A Theoretical and Experimental Study on a Stratified
The major drawback of these experiments was the fact that it Downdraft Biomass Gasifier, Ph.D. Thesis, University of Melbourne,
Melbourne, Australia, 2004.
was not possible to integrally replace the argon by nitrogen. (12) Nagai, J. M. Estudos dos Principais Processos de Ionização e
The maximum nitrogen−argon (N/Ar) mass ratio obtained Dissociação numa Descarga Luminescente em N2, Ph.D. Thesis,
was 1.9. Above this ratio, the torch became unstable and Aeronautical Institute of Technology, São José dos Campos, Brazil,
quenching was inevitable. Another drawback was the 2004.
accumulation of carbon inside the torch reactor quartz tube, (13) Gomes, M. P. Microplasmas em Equilibrio ́ de Excitação, Ph.D.
preventing the microwave from reaching the gas molecules and Thesis, Aeronautical Institute of Technology, São José dos Campos,
also leading to quenching. Perhaps both problems can be Brazil, 2011.
solved by increasing water concentration in the gas upstream (14) National Institute of Standards and Technology (NIST).
from the reactor. Such an increase in water doping can be Atomic spectra data base, http://physics.nist.gov/cgi-bin/ASD/lines1.
pl (accessed Aug. 18, 2009).
achieved by raising the liquid water content in the tar solution
(15) Yong, Y. C.; Lee, J. S.; Shin, D. H.; Kim,Y. J.; Shin,Y. U.; Lee, J.
or injecting water steam. It is interesting to remember that the B.; Uhm, H. S. Plasma Burner Enlarged by Coal Injection into
water concentration in this experiment was far below the typical Microwave Plasma. IEEE Trans. Plasma Sci. 2011, 39 (9), 1829−1833.
water concentration in the gasifier-eluded gases. (16) Uhm, H. S.; Cho, S. C.; Hong, Y. C.; Park, Y. G.; Park, J. S.
■
Destruction of Dimethyl Methylphosphonate Using a Microwave
AUTHOR INFORMATION Plasma Torch. Appl. Phys. Lett. 2008, 1−4.
Corresponding Author
*Tel.: (+55 91) 3201 8343. Fax: (+55 91) 3223 9534. E-mail:
mfmn@ufpa.br.
Notes
The authors declare no competing financial interest.
1181 dx.doi.org/10.1021/ef301399q | Energy Fuels 2013, 27, 1174−1181
Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.
Alternative Proxies: