Journal of Minerals & Materials Characterization & Engineering, Vol. 8, No.4, pp 249-259, 2009
Printed in the USA. All rights reserved
Petroleum Coke Particle Size Effects on the Treatment of EAF Dust through
Microwave Heating
Xiang Sun, Jiann-Yang Hwang*, Xiaodi Huang, and Bowen Li
Department of Materials Science and Engineering
Michigan Technological University, MI 49931
*Corresponding Author, contact:
Phone: 906-487-2600, Fax: 906-487-2934
EAF dusts were mixed with petroleum coke and irradiated together under microwave for a
pyrometallurgical treatment. It was found that particle size of the reducing agent played an
important role in affecting the reduction degree. Both zinc removal and metallization degree
increased with the decreasing of the coke particle size. By changing both microwave time and
carbon addition, optimal zinc removal at 99.23 % and metallization at 100 % can be achieved
with 15 minutes microwave irradiation and 20 wt% carbon addition.
Key Words: EAF dust, microwave processing, zinc recovery, metal extraction
EAF (electric arc furnace) dust is formed under minimill steelmaking operations due to the high
processing temperature (around 1600
C). Certain metals such as Zn, Pb, Cd, Na, Mn and Fe are
volatilized and oxidized, then condensed or mechanically carried over and finally collected as
appeared as dust form. Because it contains water leachable heavy metals such as lead, chromium
and cadmium, EAF dust was categorized as hazardous waste K061 under EPA regulation in
1980 [1]. Management of EAF dust becomes a major economic problem in US minimill
steelmakers due to the environmental concern. Several technologies have been developed and
some of them have been constructed in commercial scale. So far most processes are concerned
250 X. Sun, J.-Y. Hwang, X. Huang, and B. Li Vol.8, No.4
with pyrometallurgical and hydrometallurgical ways to recover valuable metals such as zinc or
iron. However, both technical and economical problems are still remaining.
Microwave interacts with materials at molecule level and heat can be generated inside materials.
The energy consumption and process time can be highly reduced. Due to the great and unique
advantages, microwave has been applied to numerous industries and still attracts lots of research
interests [2-4]. Microwave treatment of EAF dust has also been studied and zinc removal from
the hazardous waste has been proved feasible [5-6].
In metal extraction, the carbon reducing agent plays an important role. Carbon particle size is
usually the major concern. In microwave heating, the carbon is directly interacting with the
electromagnetic waves. Carbons at different particle sizes may have various heating
consequences thus leading to different extraction results. This is not common when conventional
heating methods are utilized. Moreover, in order to utilize such advanced technique into real
industry, inexpensive reductant is a critical factor to reduce the total cost. Petroleum coke is an
inexpensive carbon source, thus it could be an ideal reductant source for industrialization of
microwave metal extraction. Therefore, a study on the coke particle size effects in microwave
treatment of EAF dust is required.
This paper presents a laboratory investigation of processing EAF dust through microwave
heating, with focus on the carbon particle size effects on the reduction degree and associated
2.1. Experimental System
Experiments were conducted in a modified microwave oven. The details of setup have been
described in our previous publication [6]. EAF dusts were mechanically mixed with petroleum
cokes, and then put into a microwave transparent clay crucible. A crucible holder with a pipe
connected was used to rivet the crucible and direct the outgoing gases or particles. A glass funnel
was loosely put right on top of the pipe and connected to the particle collector. A paper filter
which has a pore size of 2.5 µm was fixed inside the collector and used to hold up the solidified
particles. The metal or compound vapors from the crucible are sucked by a pump, solidified, and
trapped in the collector.
2.2 Characterizations
Inductively coupled plasma (ICP) was utilized to determine the chemical compositions. The
dissolution of all the elements was conducted by using combination of HCl and HNO
Vol.8, No.4 Petroleum Coke Particle Size Effects on the Treatment of EAF 251
Metallic iron was determined by ICP after dissolving samples in 0.5 N CuSO
solutions instead
of acid mixtures [7] based on the following reaction:
CuFeSOCuSOFe +→+
The solution was heated until boiling for 30 minutes on a hot plate. Continuous stirring is
necessary in order to prevent reduced copper precipitates to form a Cu thin film around the
metallic iron particles which may prevent these particles from further oxidation. The later
solution was filtered and diluted, then analyzed by ICP.
Elemental analyzer NA1500 (Fisons Instruments Inc) were used to determine the total carbon
contents. The phase compositions were determined by getting X-ray diffraction (XRD) patterns
on a Scintag XDS2000 powder diffractometer with a graphite monochromoter using Cu Kα
radiation. Scanning electron microscope (SEM) JEOL JSM-6400 with EDS analyzer was utilized
to study morphology and microstructures.
Processing temperature was determined by using shielded thermocouples. However, there are
studies [8-9] showing that with present of thermocouple in microwave field, the local distribution
of electromagnetic field can be highly distorted and abruptly changed. This may lead inaccurate
reading and processing problems such as local breakdown or thermal runaway. Therefore, in this
study, we measured the temperature by immediately inserting the thermocouple into the sample
right after the time up of the microwave power. The heat loss to the surroundings was slowed by
the refractory, so obtained temperature is not too far from the real number.
2.3 Raw Materials
Petroleum coke is utilized as the carbon source for this study. Since it is a by-product from the
petroleum refineries, it is a very inexpensive material. Samples from Mid-Continent Coal and
Coke Co. were obtained for this study. Table 1 shows the composition of the material. It has
92.6 wt% carbon and 6.7% of sulfur and tiny amount of iron, vanadium, nickel, etc. The sample
was grounded and sequentially screened to +16, 16-40, 40-60, 60-100, 100-140, 140-200, 200-
270, and -325 meshes. Carbon addition was calculated by carbon percentage.
Table 1 Chemistry of obtained petroleum coke
92.6 6.7 0.61 1323 1347 227 110 93 307
252 X. Sun, J.-Y. Hwang, X. Huang, and B. Li Vol.8, No.4
Raw EAF dusts were received from Steel Dynamics, Inc. The X-ray diffraction pattern, as in
Figure 1, showed the major phases in the dust are ZnFe
(Franklinite), ZnO (Zincite) and KCl
(Sylvite). It also presents that most of zinc and iron are in oxides form.
Figure 1 XRD pattern of EAF dust
The morphology of the EAF dust was observed from SEM analysis. As shown in Figure 2, the
dust generally contains very tiny (less than 2 µm) spheroids, some of which agglomerate into
relative large particles (10-100 µm). Figure 2(a) and (b) illustrate composition variations of
particles in different sizes. The small particles mainly consist of ZnFe
and Fe
, which fill
about 80-90% of the whole dusts. Medium size particles are metal oxides or silicates. The big
particles are mostly Fe-enriched silicates or oxides and there are fine oxides particles attached on
them. In general, EAF dust contains mainly of ZnFe
, FeO, ZnO, minor amount KCl,
Fe-Al-Ca-Zn-Mg oxides, and trace amount of Fe
and various silicates. However, nearly none
of these compounds was found to be pure in composition. Element substitution for each other is
common in these compounds [10]. For example, in a typical ZnFe
particle, 1.33 % Ca, 1.2 %
Mg, 1.78% Mn was observed to substitute for Zn and Fe. These substitutions indicate the
complexity of the dust compositions.
The carbon-dust mixtures were irradiated by microwave so that heats were generated inside the
materials and a homogeneous heating was expected. When the temperature was favorable,
gasification of carbons and carbon monoxide began to occur and sequentially reduction of
Vol.8, No.4 Petroleum Coke Particle Size Effects on the Treatment of EAF 253
compounds in the dust got started. At even higher temperature, reduced metallic zinc would
vaporize out.
Figure 2 General morphology of EAF dust
(a) 1, 2 – Fe-Ca-Mg oxides, 3 – Ca-Fe-Zn oxides, 4 – ZnFe
, 5 – FeO, 6 – Fe-Ca silicates, 7 –
Fe-Ca oxides, 8 – CaO
(b) 1 – ZnFe
, 2 – Fe
, 3 – Fe
, 4 – Fe-Zn-Al silicates, 5 – metallic Fe
Here, our concerns are focused on iron and zinc since they are the most abundant valuable
elements in the dust. An idea outcome from the process is the perfect extraction and separation
of Fe and Zn. The metallization degree and zinc removal are calculated by following equations:
Where, is the weight of metallic iron and is the total iron content in the final
product; is the weight of total zinc in the product and is the total zinc content in
the original dust.
3.1. Carbon Size Effects
To study the carbon size effects, each test was performed under the same carbon addition (15
wt%), microwave power (1.1 kW) and microwave time (15min). The results of zinc removal and
iron reduction degree (metallization) are shown in Figure 3. The zinc removal and metallization
degrees both increased with decreasing of petroleum coke particle size. At coke particle size less
than 44 µm (-325 mesh), zinc removal and metallization reached the highest degree, which were
92.73% and 99.59%, respectively.
254 X. Sun, J.-Y. Hwang, X. Huang, and B. Li Vol.8, No.4
Figure 3 Effects of petroleum coke size on the zinc removal (left) and metallization degrees
(right), MW power 1.1kW, processing time 15min, coke addition 15 wt%
The original dusts have most particles smaller than 2 µm. As the coke particle size went smaller,
the surface areas for chemical reactions became larger. This is believed as a major cause for the
better reduction degree at finer coke particle size.
On the other hand, heat conduction may be another important issue. Table 2 shows the sample
temperature of EAF dust and petroleum coke particles at various sizes after being radiated by
microwave for 5 minutes. EAF dust couples with microwave very well and can be heated very
fast due to the presence of ZnFeO4, which is a magnetic oxide [11]. However, petroleum coke is
not a good microwave absorber and even for smaller particle size, the absorption doesn’t have
any significant improvement. When the mixture was heated by microwave, EAF dust absorbed
most of the energy and coke got temperature increase by heat conduction from the dust. Smaller
particles have larger contact areas for heat conduction which can cause faster gasification of
carbon. This is obvious from the results of residual carbon analysis. Such observation is unique
for petroleum coke since graphite and coke have much better interaction with microwave [6].
Therefore, in order to use inexpensive carbon source for treatment of EAF dust, smaller coke
particles should be used to have good treatment outcome.
Table 2 Temperature after 5min microwave heating (20 g sample, power 1.1 kW)
From Table 3, there are significant amount of unburned carbon left for bigger coke particles after
processing. And the percentage of carbon left decreased with the decreasing of carbon size. So
EAF dust
+16 mesh
pet coke
60 mesh
pet coke
200 mesh
pet coke
325 mesh
pet coke
C 231
C 232
C 202
C 221
Vol.8, No.4 Petroleum Coke Particle Size Effects on the Treatment of EAF 255
both larger reaction area and better heat transfer contributed to the optimum removal and
metallization for the smallest particle size of carbon.
Table 3 Residual carbon after processing (wt %)
Figure 4 shows some XRD patterns of the reduced solids in the crucible. The residuals contain
Fe in majority, minor amount of FeO, ZnS, KCl, and Ca
, and trace amount of
. As the coke size decreased, the FeO peak is diminishing, which proves the better
metallization degree. The amorphous carbon left in the +16 mesh and 16-40 mesh fractions
indicates the existence of significant amount of unburned coke in the coarse fractions. Also
significant ZnO and ZnS peaks were found in mixtures of big particles of carbon. Petroleum
coke is rich (6.7 wt%) in sulfur. So it is not surprised to see ZnS compounds formed. However,
as the particle sizes became smaller, peaks of these zinc compounds diminished. When the
reaction chamber is under a non-reducing atmosphere, as what happens in samples with bigger
coke particles, fewer ZnO could be reduced and also part of the reduced Zn could be reoxidized.
Calcium and magnesium silicates were observed to be the most abundant slag phases.
Figure 4 XRD patterns of the reduced mixtures at different carbon sizes
+16 mesh 16-40
7.4 6.3 2.4 1.6 <1 <1 <1 <1
256 X. Sun, J.-Y. Hwang, X. Huang, and B. Li Vol.8, No.4
3.2. Processing Time and Carbon Addition Effects
Even though almost 100% metallization degree was achieved at the smallest coke size, there was
still significant percentage of zinc residual in the products, which showed that the carbon for
reduction is not enough. So carbon additions (coke size -325 mesh) as well as microwave time
were varied to study the effects.
As shown in Figure 5, the reduction degrees for both zinc and iron improved consistently with
the increasing of microwave processing time and carbon addition. The reduction degree of iron is
always higher than that of zinc, which indicates the faster reaction rates for iron oxides than zinc
oxides. Iron oxides have very good absorption of microwave and can be heated up rapidly.
However, zinc oxides are very poor microwave absorbers [12]. In the processing, iron oxides
absorbed most of the microwave energy, thus the reduction of the iron oxides were favored. Part
of zinc in the ferrites is reduced together with iron. But the zinc in its own oxide and sulfide
forms is reduced at lower rate. Near complete removal of zinc and reduction of iron oxides can
be achieved at 20 wt% carbon addition and 15 minutes microwave time.
Figure 5 Zinc removal and metallization degree as a function of microwave time and carbon
addition (■)10 wt%, (●) 15 wt% (▲) 20 wt%, microwave power 1.1 kW
Vol.8, No.4 Petroleum Coke Particle Size Effects on the Treatment of EAF 257
3.3. Products
Two kinds of products have been obtained. One is the iron rich residual in the crucible, and the
other is the collected particles.
A chemical analysis of the iron rich residual from the 20% carbon addition and 15 minutes
microwave radiation run is shown in Table 4. The residue contains 69.5% metallic iron, 69.8%
total iron, 5.5% calcium, 2.9% manganese, 2.9% magnesium, and less than 0.5% of zinc,
chromium, lead, copper, sodium. Apparently, metallization of iron is near 100%.
Table 4 Typical chemical analysis of iron rich residue (20 wt% carbon addition, MW 15min)
Element Fe(M) Fe(T) Zn Cr Ni Cd
Wt % 69.49 69.84 0.138 0.29 0.084 <0.01
Element Pb Si Mn Mg Cu Ca Na
Wt % 0.129 0.997 2.839 2.898 0.237 5.533 0.44
SEM pictures show the morphology of the residuals. As in Figure 6, metallic iron is the major
phase and small particles agglomerated together forming big spheroids. About 2-3 at% Mn has
been found in most metallic iron particles. The residual mixture also contains minor amount of
Ca-Mg silicates, trace amount of carbides, sulfides and unburned carbon. The silicate rich slag is
surrounding the iron particles and the separation between them is obvious. Some molten phases
have formed and the residue has partly sintered.
Figure 6 SEM picture of residual mixture (carbon addition 15 wt%, MW 15min)
1 – metallic Fe, 2 – Mg-Ca silicates, 3 – Ca
, 4 – CaS, 5 – residual carbon, 6 – Ca-Al
silicates, 7 – Mg-Ca silicates and sulfides
258 X. Sun, J.-Y. Hwang, X. Huang, and B. Li Vol.8, No.4
A XRD pattern (Figure 7) of the collected particles presented that the most abundant phase is
metallic zinc. KCl, CrCl
and metallic Pb are in minor amount. Very small peak of ZnO is also
Figure 7 XRD analysis of the collected particles
Microwave has been tested for processing of electric arc furnace dusts in laboratory scale.
Particle size of reducing agent has potential effects on the reduction degrees. Both zinc removal
and metallization degrees have been found to increase with decreasing of the petroleum coke
particle size. Also by changing the microwave time and fixed carbon addition, optimal yield has
been achieved. 99.23 % zinc removal and 100 % metallization have been obtained at 15 minutes
microwave time and 20 wt% carbon addition. Two kinds of products, an iron rich residual in the
crucible and a metallic zinc rich product containing particles collected from condensed vapors
have shown the excellent separation and potential recycling possibility.
The authors would like to thank Steel Dynamics, Inc. for providing dust samples and Kaikun
Yang for SEM analyses.
[1] Sofilić, T., Rastovčan-Mioč, A., Cerjan-Stefanović, Š., Novosel-Radović, V., and Jenko, M.,
2004, “Characterization of steel mill electric-arc furnace dust.”
J. Hazard. Mater.,
Vol. 109,
pp. 59-70.
Vol.8, No.4 Petroleum Coke Particle Size Effects on the Treatment of EAF 259
[2] Hwang, J.Y., and Huang, X.,
Advanced Processing of Metals and Materials, Volume 5—New,
Improved and Existing Technologies: Iron and Steel and Recycling and Waste Treatment, ed.
F. Kongoli and R.G. Reddy (Warrendale, PA: TMS, 2006), “New Steel Production
Technology with Microwave and Electric Arc Heating.” pp. 251–261.
[3] Clark, D. E., Folz, D. C., and West, J. K., 2000, “Processing materials with microwave
Mater. Sci. Eng., A
, Vol. 287, pp. 153-158.
[4] Shi, S., and Hwang, J. Y., 2003, “Microwave-assisted wet chemical synthesis: advantages,
significance, and steps to industrialization.”
J. Miner. Mater. Character. Eng.
, Vol. 2, pp.
[5] Ghoreshy, M., and Pickles, C.A., 1995, “Microwave processing of electric arc furnace dust.”
Electr. Furn. Conf. Proc.
, Vol. 52, pp. 187-196.
[6] Sun X., Hwang, J. Y., and Huang X., 2008, “The microwave processing of electric arc
furnace dust.”
, Vol. 60, pp. 35-39.
[7] Xu, Z., Hwang, J. Y., Greenlund, R., Huang, X., Luo, J., and Anschuetz, S., 2003,
“Quantitative determination of metallic iron content in steel-making slag.”
J. Miner. Mater.
Character. Eng.,
Vol. 2, pp. 65-70.
[8] Thompson, K., Booske, J. H., Cooper, R. F., Gianchandani, Y. B., and Ge, S., 2001,
“Temperature measurement in microwave-heated silicon wafers.”
Ceram. Trans.
, Vol. 111,
pp. 391-398.
[9] Pert, E., Carmel, et al., 2001, “Temperature Measurements during Microwave Processing:
The Significance of Thermocouple Effects.”
J. Am. Ceram. Soc.
, Vol. 84, pp. 1981-1986.
[10] Chen, T. T., Dutrizac, J. E., and Owens, D. R., 1998, “Mineralogical characterization of
EAF dusts from plain carbon steel and stainless steel operations.”
Waste Process. Recycl. III
pp. 511-525.
[11] Rogers A., and Alexander, J., 1920,
Industrial Chemistry
, 3
ed., D. Van Nostrand, New
[12] Saidi, A., and Azari, K., 2005, “Carbothermic reduction of zinc oxide concentrate by
J. Mater. Sci. Technol.
, Vol. 21, pp. 724-728.