GCI
TECH NOTES©
Volume
12, Number 2
A Gossman Consulting, Inc.
Publication February 2007
There is a growing level of concern
about mercury emissions from cement kilns and interest in the industry
in developing cost effective options for controlling these emissions.
Cement plants have a wide range of mercury inputs and resulting
emissions because of the wide variety of raw materials and fuels used
in the process. Further, the current level of mercury emission control
at cement plants varies from 0% to as high as 95% using existing
particulate control systems. This is the third in a new series
of GCI TechNotes that will examine this issue.
Mercury
emissions are regulated based on concern for mercury
entering the food chain and bioaccummulating to significant levels that
could
impact people eating fish. The following is a brief review of the
factors that
impact the issue of controlling mercury emissions from modern cement
kilns.
It should be kept in mind that mercury control technologies are under active development around the world as pressure mounts to reduce the total anthropogenic source of mercury to the mercury cycle. Most of this development has been focused on controlling mercury emission from coal fired electrical power utility boilers. Not all of the observations made regarding mercury emissions and controls for those systems can be directly applied to cement kilns. Further, the unique operating environment inside a cement kiln may present innovative and cost effective control methods for cement kilns that are impossible or impractical to apply to coal fired boilers.
Review of Control Options
During combustion, the mercury (Hg) in coal and other fuels is volatilized and converted to elemental mercury (Hg0)
vapor in the high temperature regions of the cement kiln. As the flue
gas is cooled, a series of complex reactions begin to convert Hg0 to ionic mercury (Hg2+) compounds and/or Hg compounds (Hgp)
that are in a solid-phase at flue gas cleaning temperatures, such as
HgO and HgS, or Hg that is adsorbed onto the surface of other
particles. The presence of chlorine gas-phase equilibrium can favor the
formation of mercuric chloride (HgCl2) at system
temperatures depending on the competition of alkali-chloride reactions.
The presence of significant levels of sulfide in the form of iron
sulfide (pyrite) in some cement kiln raw feeds may favor the formation
of HgS. However, Hg0 oxidation reactions are
kinetically limited and, as a result, Hg can enter the air pollution
control device(s) as a mixture of Hg0, Hg2+, and Hgp with varying ratios depending on the conditions in the kiln system. This partitioning of Hg into Hg0, Hg2+,
and Hgp is known as mercury speciation, which can have considerable
influence on selection of mercury control approaches. For this reason
it is critical that assessing the range of operating conditions and the
resulting speciation of Hg in any given cement kiln be the first step
in determining the most effective control technology. A control
technology that works on one kiln system cannot be assumed to be
effective on another.
It is also critical to understand the role of CKD recycling in this
process. Recycling CKD back into the kiln system can revolatilize
mercury that has been captured, converting it from the particulate form
to one of the more volatiles forms such as the chloride.
Some control of mercury emissions from cement kilns is currently
achieved via existing controls used to remove particulate matter (PM),
sulfur dioxide (SO2), and nitrogen oxides (NOx). This includes capture of Hgp in PM control equipment and soluble Hg2+
compounds in wet scrubber systems. Available data on electric utility
boilers also suggest that use of selective catalytic reduction
(SCR) NOx control enhances oxidation of Hg0
in stack gasses and results in increased mercury removal in wet
scrubbers. It is unlikely that the same would be true of SNCR NOx
control systems on cement kilns given the operating temperature of
these systems. That said, the presence of excess ammonia in the stack
gasses could impact available chlorides for reacting with and forming
mercuric chloride.
There are three broad approaches to mercury control: (1) activated
carbon injection (ACI), (2) multi-pollutant particulate control, in
which Hg capture is enhanced in existing/new PM control devices, and
(3) multi-pollutant wet scrubber control, in which Hg capture is
enhanced in existing/new SO2, and NOx
control devices. Relative to these three approaches, this paper
describes currently available data, limitations, estimated potential,
and research and development needs.
Activated Carbon Injection Control of Mercury Emissions
ACI has the potential to achieve moderate levels of mercury control.
The performance of activated carbon is related to its physical and
chemical characteristics. Generally, the physical properties of
interest are surface area, pore size distribution, and particle size
distribution. The capacity for mercury capture generally increases with
increasing surface area and pore volume. The ability of mercury and
other sorbates to penetrate into the interior of a particle is related
to pore size distribution. The pores of the carbon sorbent must be
large enough to provide free access to internal surface area by Hg0 and Hg2+
while avoiding excessive blockage by previously adsorbed reactants. As
particle sizes decrease, access to the internal surface area of
particle increases along with potential adsorption rates.
Carbon sorbent capacity is dependent on temperature, the concentration
of mercury in the flue gas, the flue gas composition, and other
factors. In general, the capacity for adsorbing Hg2+ will be different than that for Hg0.
The selection of a carbon for a given application would take into
consideration the total concentration of mercury, the relative amounts
of Hg0 and Hg2+, the flue gas
composition, and the method of capture (electrostatic precipitator
(ESP), or baghouse). An important factor for some cement kilns will be
the levels of hydrocarbons and the need to account for their sorption
on to the carbon reducing the capacity of the carbon to adsorb mercury.
In addition, bench-scale research shows that high SO2
concentrations diminished the adsorption capacity of activated carbons.
Both of these issues could prevent ACI from being an effective control
on some cement kilns.
There has been only limited testing of ACI on low concentration mercury
gas streams as are typical of cement kilns. Most of this work has been
done on power plant boilers achieving control efficiencies of 25-95%
depending on the type of coal being burned and a wide number of other
factors. In many cases these plants already had some mercury control
via the particulate control systems in place and enhanced control via
ACI was as low as a 10% improvement.
ACI has the further disadvantage of requiring the disposal of the
mercury contaminated spent carbon. Whether the carbon is cleaned and
reactivated for reuse or disposed of, the ultimate fate of the mercury
needs to be assessed to insure that the mercury will not be
reintroduced into the global mercury cycle through some other means.
Mercury Emission Control via Control of Particulates in ESPs or Baghouses
ESPs and/or baghouses (fabric filters) can be an effective control for
mercury from cement kilns if two critical conditions are met. First,
the mercury must be in the particulate form. This may occur naturally
in the system or may require reagents added to the right point in the
process to oxidize or catalyze the oxidation of the Hg to HgS and/or
HgO. One US wet process plant has demonstrated 95% control of mercury
emission through their existing ESP system. High levels of pyrite in
their raw materials may be a factor in producing this relatively high
level of control. (Theoretically, efforts to reduct SOx
emissions by controlling pyrite in the raw feed could increase mercury
emissions.) One experimental system (not in a cement kiln) uses UV
light to shift the oxidation state of mercury.
The second critical step in the process is to remove a portion of the
mercury-containing dust from the air pollution control system in such a
way as to maximize the mercury removal and not place that portion of
the dust back into the kiln system. In older wet process plants this
has been done routinely by removing the dust from the final one or two
stages of the ESP systems. Precalciner plants with in-line raw mills
have a more complex scenario to consider. For plants that currently
recycle all of their cement kiln dust, the mercury is simply returned
to the system and recirculates until the concentration gets high enough
that a portion is emitted out the stack in some form. It is critical to
break this cycle. Without breaking this cycle the speciation of the
emissions may have limited or no meaning relative in determining the
most effective control technology.
A program to speciate mercury in dust samples taken during raw mill on
and off conditions should be a first step in characterizing an
operation to see if particulate control of mercury emissions is a
viable option. It needs to be kept in mind that under current normal
operations where all CKD is recycled back into the system that the
mercury concentration in stack gasses and in the dust will be highly
dynamic. The system probably never reaches a steady-state operation
relative to mercury input and output. (This implies that there is no
way to accurately determine mercury emissions with a stack test.)
Modeling of cement kilns with in-line raw mills suggests that removing
a small portion of the captured dust when the raw mill is operating
(and possibly when it is not on line as well) can break the recycle
loop and may control mercury emissions with efficiencies in excess of
90% depending on speciation and a number of other factors including
baghouse blowback cycles, baghouse (or ESP) operating temperatures,
types of bags, etc. With the raw mill operating there is likely a very
high level of sorption of mercury onto particles in the raw mill and in
the baghouse (or ESP) – one set of tests on a precalciner showed
sorption efficiencies of 98.5%. It has been typical to operate
baghouses at higher temperatures when the raw mill is down and in the
just mentioned case efficiencies dropped to 90%. This drop in
efficiency is likely to have been due to the increase in the baghouse
operating temperature from 100 °C to 175-200 °C.
Spray Tower/Wet Scrubbing of Mercury Emissions
Wet spray tower/slurry systems remove gaseous SO2 from emissions by absorption. For SO2 absorption, gaseous SO2 is contacted with a caustic slurry, typically water and limestone or water and lime. Gaseous compounds of Hg2+ are generally water-soluble and can absorb in the aqueous slurry of a wet scrubber system. However, gaseous Hg0 is insoluble in water and therefore does not absorb in such slurries. When gaseous compounds of Hg2+
are absorbed in the liquid slurry of a wet system, the dissolved
species are believed to react with dissolved sulfides from the flue
gas, such as H2S, to form mercuric sulfide (HgS); the HgS precipitates from the liquid solution as sludge.
The capture of mercury in wet scrubbers is likely dependent on the relative amount of Hg2+
in the inlet flue gas and on the PM control technology used. Electric
utility boiler data reflected that average mercury captures ranged from
29 percent for one unit burning subbituminous coal with an ESP plus wet
scrubber to 98 percent in a unit burning bituminous coal with a fabric
filter baghouse plus a wet scrubber. The high mercury capture in the
fabric filter baghouse plus wet scrubber unit was attributed to
increased oxidization and capture of mercury in the baghouse followed
by capture of any remaining Hg2+ in the wet scrubber. For cement plants with SOx
scrubbers this has particular potential. A system of bleeding the APCD
particulate control system followed by a scrubber system for SOx that coincidentally captures HgCl2 may provide very high levels of mercury emission control on cement plants with the right chemistry.
Conclusion
While ACI and wet scrubbing may provide control of mercury emissions
from cement kilns, the lowest cost option appears to be the use of the
existing particulate control system in conjunction with a small bleed
of dust from the primary air pollution control system. Plants that have
an ESP, may find that the dust in the final stages of the ESP is even
more enriched in mercury and that this simplifies the process of
creating a “break” in the mercury recycle loop. For
example, if it is found that this dust represents 1% of the total feed
to the kiln system and is enriched to a factor of 100 times the average
level of mercury in the system relative to raw feed; removal of that
dust would effectively remove all the mercury from the system. This
dust could then be sent to the finished cement blending silos with no
appreciable impact on product quality. Investigation of the speciation
and enrichment of mercury in the dust being captured in various stages
of the ESP or bagouse with the raw mill both on and off is recommended
as the first step in developing a dynamic model and from that a mercury
control strategy for any cement plant wishing to reduce mercury
emissions.
References
CRC Handbook of Chemistry and Physics 70th Edition (1989). Ed. Weast, Robert C., Ph.D., Florida: CRC Press, Inc.
Merck Index Twelfth Edition, The (1996). Ed. Budavari. New Jersey:
Merck Research Laboratories, Division of Merck & Co., Inc.
Control of Mercury Emissions from Coal-Fired Electric Utility Boilers:
An Update; EPA Air Pollution Prevention and Control Division, National
Risk Management Research Laboratory, ORD: Research Triangle Park, NC,
Feb 18, 2005; http://www.epa.gov/ttn/atw/utility/ord_whtpaper_hgcontroltech_oar-2002-0056-6141.pdf.