-
Inuence of geometrical parameters of honeycomb commercial
SCR-DeNOx-catalysts on Dand SO2/SO3-conversion
T. Schwmmle a,, F. Bertsche a, Aa University of Stuttgart,
Institute of Combustion and Po
serstr. 3er-von-H
arametared inSO3-cone to en n at ca
condit ioning with SO2 due to superior acidity of active sites.
Summarising the results, due to the draw- sts, increasing wall by
SCR catalysts.All rights reserved.
Over 40% of the global electricity is produced in coal red power
stations, either by ring hard coal or lignite [1]. The total amount
ofcoal consumed globally will increase over the next years. Coal
derived ue gas consists of several pollutants, such as nitrogen
oxides (NOx), sulphur oxides (SO2, SO3), particulate matter (PM)and
trace elements e.g. mercury (Hg). Particulate matter is
use ltersremoved by dry or wet ue gas desulphurisation
(FGD).
During technical combustion, nitrogen oxides are formed by two
different mechanisms, namely by oxidation of the nitrogen content
of the fuel and depending on combustion temper- ature via oxidation
of air-nitrog en [2]. For the reduction of NOx(DeNOx), the
selective catalytic reduction process (SCR) is the most common
applied technology since it was commerciall y vended inthe 1980s
[3,4]. The overall reaction can be expresse d as:
4NO 4NH3 O2 ! 4N2 6H2O 1 Corresponding author. Tel.: +49 711 685
67760; fax: +49 711 685 63491.
Chemical Engineering Journal 222 (2013) 274281
Contents lists available at
n
w.E-mail address: [email protected] (T.
Schwmmle).back of elevated SO2/SO3-conversion, higher risk of
channel blocking and material cothickn ess cannot be considered a
reasonable strategy to enhance mercury oxidation
2013 Elsevier B.V.
1. Introduction controlle d by electrostatic precipitators or
bagho1385-8947/$ - see front matter 2013 Elsevier B.V. All rights
reserved.http://dx.doi.org/10.1016/j.cej.2013.02.057. SO2 is
mainly Article history:Received 27 November 2012 Received in
revised form 11 February 2013 Accepted 15 February 2013 Available
online 26 February 2013
Keywords:SCR-DeNOx-catalystsDeNOx-activityMercury oxidation
SO2/SO3-conversion
A systematic study on the inuence of geometrical parameters
(pitch and wall thickness) of commercial high-dust honeycomb
SCR-DeNOx catalysts on DeNOx-act ivity, mercury oxidati on and
SO2/SO3-conver-sion is described. All catalysts had an identical
chemical composition of 0.6 wt.% V2O5. The study was con- ducted in
a laboratory micro reactor and a technical scale bench reactor and
focuses on the effect ofdifferent honeycomb geometrical parameters
on the reactions at the catalysts. The combined variation of pitch
and wall thickness showed lower DeNOx-activity at catalysts with
larger channel openings. This indicates mass transfer limitations
when the ow regimes are developed in the catalysts channels due
tothe relatively fast reaction kinetics of the DeNOx-reaction.
Results showed that the SO2/SO3-conversion islinearly depende nt on
catalysts wall thickness. Mercury oxidation increased slightly
linear with increas- ing wall thickness of the catalyst, indicating
that the reaction takes also place in the catalysts bulk because of
its slow chemical reaction kinetics in contrast to DeNOx-r eaction
being controlled by diffu- sion. The importance of ue gas
HCl-concentr ation on mercury oxidation was pointed out.
Additionally,research on the co-inuence of DeNOx-reaction and
SO2/SO3-conversion was performe d. A strong inhibi- tion of
SO2/SO3-conversion by ue gas ammonia was shown. However,
DeNOx-activity is enhanced bya r t i c l e i n f o a b s t r a c tb
IBIDEN Porzellanfabrik Frauenthal GmbH, Gamc E.ON New Build &
Technology GmbH, Alexand
h i g h l i g h t s
" Catalysts with different geometrical p" Tes ts have been
performed and comp" We investigated DeNOx-activity, SO2/" Small
pitch favou rs DeNOx-activity du" Tests showed higher mercury
oxidatioeNOx-activity, mercury oxidation
. Hartung b, J. Brandenstein c, B. Heidel a, G. Scheffknecht
a
wer Plant Technology IFK, Department Fuels and Flue Gas
Cleaning, Pfaffenwaldring 23, 70569 Stuttgart, Germany 8, 8523
Frauental, Austria umboldt-Strae 1, 45896 Gelsenkirchen,
Germany
ers have been tested.a micro reactor and bench reactor.version
and mercury oxidation.hanced external mass transfer.talysts with
increased wall thickness.Chemi cal Engi
journal homepage: wwSciVerse ScienceDi rect
eering Journal
elsevier .com/locate /ce j
-
Since combustion of coal accounts for almost 50% of the global
mercury emission s [18] and due to its high toxicity, mercury was
identied as a chemical of global concern. There are global efforts
to reduce mercury emissions from coal red power plants [18].Driven
by the reduction of the overall CO2-emissions in coal redpower
plants, co-ring of biomass and refuse derived fuels is analternativ
e way to CCS-tech nology to reduce CO2-emissions [19].
However , standard commercial high-dust SCR-DeNOx -catalysts
have not initially been designed for increased mercury oxidation
over a long time at various fuels. Addition ally, high-dust
SCR-
by kneading and extrusion followed by calcination in full
scale
ineering Journal 222 (2013) 274281 275Ammonia (NH3) is therefore
injected upstream of the catalyst unit and reacts with NOx at the
catalyst to nitrogen (N2) and water (H2O). SCR-catalyst s are
design ed either as homogeneous extruded ceramic honeycom bs, plate
catalysts or corrugated catalysts. Vari- ous catalyst materials
have been investi gated in the last decades [5]. Today, the most
common carrier mater ial for power plant appli- cation is TiO 2 in
its anatase modication and vanadia (V2O5) as ac- tive phase [6].
Certai n promoters, such as WO3 or MoO 3, are added in order to
increase the therma l stability of the TiO 2 phase [7].
As an undesired side reaction, sulphur dioxide is heteroge-
neously oxidized to sulphur trioxide at the SCR catalysts, which is
called SO2/SO3-conversion:
SO2 12 O2 ! SO3 2
The rate of this reaction is linked to the vanadia content of
the cat- alyst [5]. There are various reactions of SO3 with other
ue gas com- ponents, causing corrosion or deposit s downstre am of
the SCR-unit.In SCR-DeN Ox-cataly sts, WO3 is introduce d to
suppress the SO2/SO3conversio n [4].
Depending on their origin, coals contain 0.02 up to 10 mg/kg
mercury [8]. At high temperatures in the combusti on
chamber,mercury is released in its elemental form Hg0 [9]. When ue
gas cools down, gas-phase transformat ions occur and elemental mer-
cury is homogeneousl y or heteroge neously oxidized to Hg2+. Due to
adsorption processes , mercury is partially bound on y ash par-
ticles (HgP). Elemental mercury is considered as inert and is
almost insoluble in water. In ue gases of combustion processes, the
dom- inant oxidized mercury compound is HgCl 2, as it was shown
bytheoretical calculations performed by Martel [10]. The overall
oxi- dation reaction of elemental mercury in ue gas can be
described as:
2Hg0 4HCl O2 $ 2HgCl2 2H2O 3The impact of SCR-DeN Ox-cataly sts
on the oxidation of mercury was investigate d by several authors in
lab- and full scale measure- ments [9,11,12 ]. They demonstrat e
consistently that the oxidatio nof mercury is enhanced in the
presence of SCR-cata lysts. The results were highly dependent on
coal compositio n, especially on haloge ncontent (Cl, Br), but also
on sulphur content. Catalyst deactivatio nstate plays a key role in
mercury oxidation. There are various reac- tion-pat hways discussed
for mercury oxidatio n at SCR-cata lysts.Niksa and Fujiwa ra [13]
and Hong et al. [14] proposed EleyRideal reaction mechani sms, in
which HCl adsorbs on the active sites ofthe cataly st surface,
competi ng with ammonia. Either in the gas phase or as weekly
adsorbe d species , mercury reacts with the adsorbed HCl. Another
possible mecha nism is the Deacon process,proposed by [11] and
described as:
4HCl O2 $ 2Cl2 2H2O 4
Hg0 Cl2 $ HgCl2 5In this possible reaction pathway for mercury
oxidation atSCR-cata lysts, Cl2 is produce d by the reaction of HCl
with cataly stactive sites (Eq. (4)). Highly reactive Cl2
subsequent ly oxidises Hg0
in the gas phase (Eq. (5)). It was shown by [15] that the direct
oxidation of elemental mercury with Cl2 and Br2 in the gaseous
phase is faster than the indirect reaction of Eq. (3).
Due to its physical and chemical properties [16], HgCl 2
adsorbson y ash and can be absorbed in wet ue gas desulphurisati
onunits [9,17], leading to lower overall mercury concentrations
inthe gas phase. Therefore, the combinati on of SCR-DeNO
x-catalyst,
T. Schwmmle et al. / Chemical EngESP and wet-FGD, which is a
common set-up for installed air pollution control devices in
Western Europe, forms an ideal arrangement with the co-benet of
mercury removal.monolith size (150 mm 150 mm) by IBIDEN Porzellan
fabrik Frauenthal GmbH, Austria. All catalysts had the same
chemical compositi on, which was veried by surface and bulk
XRF-analysis.Geometri cal properties of honeycomb catalysts can be
character- ized by pitch, wall thickness and length. Fig. 1 shows a
honeycomb catalyst sample, with the homogeneous catalyst wall (w)
and the clear width (d) of the ow channel. The catalyst pitch is
equal toone clear width and one wall thickness. A combined
variation ofpitch and wall thickness has been performed, in order
to get anoverview on both: channel opening and wall thickness of
the cat- alyst. In Table 1, the dimensions of the tested catalysts
are listed:
The catalysts are named according to their pitch. Reference is
a7 mm pitch catalyst with 21 21 channels of the full size mono-
lith, which is the typical high-dust catalyst pitch. Variation of
pitch was performed in the range of 6 mm up to 9.7 mm and
accordingly catalysts are facing further challenges, such as
erosion and blocking by y ash [20], leading to lower efciency and
lifetime of the catalyst. Standard commerc ial honeycomb catalysts
are produced with a 7 mm pitch. Increasing the catalyst pitch
athoneycomb catalysts would reduce the risk of blocking of the
channels. Thicker walls of the catalysts honeycomb structure would
lead to an increased lifetime of the catalyst monolith inthe ue
gas. It is assumed, that the modication of the catalysts geometri
cal parameters will also have an effect on the reactions catalysed.
In order to quantify the extent of this effect, a systematic study
at different test conditions is performed, giving an overall
insight in the most important reactions for emission reduction
oncommerc ial SCR-DeNO x-catalysts. Finally, the economic aspects
of different catalyst geometries of honeycomb catalysts are
addresse d.
2. Experimen tal method
2.1. Catalysts
In this study, the behaviou r of standard commercial honeycomb
catalysts is addressed, which are one of the major catalyst types
for power plant application. The catalysts were of
V2O5WO3/TiO2-type with a V2O5-content of 0.6 wt.%. This is a
typical value for high-dust catalysts in coal-red power plants
located downstream of the boiler [5,20,21]. Honeycomb catalysts
have been produced Fig. 1. Geometrical properties of honeycomb
catalysts.
-
wall thickness was varied between 0.8 mm and 1.4 mm, covering
the whole reasonable range for high-dust application.
2.2. Test setup
Investigations on the reactions at the catalyst have been per-
formed at two different test rigs: The micro reactor and the bench
reactor, both constructed and operated accordin g to the guideline
for testing of DeNOx catalysts [22]. Fig. 2 describes the general
set- up of the reactors.
In both test rigs, ue gas is doped with the relevant compo-
nents, which are added from gas bottles or by evaporati on of
liquid solutions of these compounds . Table 2 summari zes the test
condi- tions. Water content is adjusted either by evaporation or by
con- densation. The gas mixture is heated up in the
reactor.Thermocoupl es up- and downstream of the catalyst ensure a
con- stant temperature of 380 C. Samples are taken at ports up- and
downstream of the catalyst and analysed as described below. For
Table 1Dimensions of the tested catalysts.
Pitch (mm) Wall (mm)
P6 6.0 0.8 P6.7 6.7 0.9 RC 7.0 0.9 P8.2 8.2 1.0 P9.2 9.2 1.2 P10
9.7 1.4
276 T. Schwmmle et al. / Chemical EngineeTable 2Test
conditions.the operation of the test reactors and testing of the
catalysts, the linear velocity (LV), describin g the ue gas ow rate
divided bythe catalyst surface area exposed to ow, and area
velocity (AV),which is the quotient of ue gas ow rate and geometric
surface
Fig. 2. Test reactor design.Micro reactor Bench reactor
LV (m/s @ STP,wet) 1 2.5 AV (m/h @ STP,wet) 27.5 1018Channels Up
to 9 Up to 625 Flue gas ow rate (m3/h @ STP) 1.2 150 Temperature
(C) 380 380 O2 (vol.%) 4 5CO2 (vol.%) 15 14N2 Balance Balance H2O
(vol.%) 7 6NO (ppmv; STP dry) 300 300 NH3 (ppmv; STP dry) 360 300
SO2 (ppmv; STP dry) 500 500 Hg (lg/m3; STP dry) 50 HCl (mg/m3; STP
dry) 10100 area of the catalyst, are of great importance . These
values are cal- culated at standard temperat ure and pressure
(STP).
2.2.1. Micro reactor In micro reactors, samples of 3 3 channels
of the full scale
monolith s are tested in laboratory in synthetic ue gas. The
reactor is constructed of glass. Flue gas ow rate is adjusted to a
LV of 1 m/s and an AV of 27.5 m/h. Due to different geometries of
the cata- lysts, the length of the catalysts has been adjusted to
meet the test condition s, resulting in catalyst samples with
differing length.
2.2.2. Bench reactor This technical scale reactor is constructed
of stainless steel and
ue gas is generated by a propane burner. In a bench reactor,
full scale catalyst monoliths of 150 mm 150 mm are tested in their
original manufactur ed dimensions . Flue gas ow rate is set to150
m3/h. Results of the bench reactor test can directly be trans-
ferred to full scale catalyst reactors and bench reactor measure
-ments are a reference for guarantee values of manufactur ers
[20].Due to safety reasons, no mercury oxidation measureme nts were
performed at this reactor.
3. Determined parameters
In order to fully cover all relevant reactions at the SCR-DeNOx
-catalysts, DeNOx-a ctivity, SO2/SO3-convers ion and mercury oxida-
tion have been researched in this study.
3.1. DeNOx-ac tivity
DeNOx activity describes the performance of the catalyst re-
lated to NOx-reduction with ammonia, considering the AV. It
isdetermined at dened and xed NH3/NO-ratios of a = 1.0 in bench
reactor and a = 1.2 in micro reactor. There is a higher NH3/NO
ratio applied at the micro reactor in order to minimise the effect
of pos- sibly unequal distribution of ammonia in the small sample
size. Atthis test condition, there is always excess ammonia in the
ue gas,showing the maximum value of NOx-reduction. DeNOx-a ctivity
iscalculated according to the rst order reaction equation in the
fol- lowing formula [22] with the AV and NOx-reduction g(a):
KNOx amh
h i AV ln1 ga 6
Nitrogen oxide concentrat ion up- and downstre am of the
catalyst ismeasured by continuou s monitor s, working on the
principle ofchemilu minescen ce [22].
3.2. SO2/SO3-conversion
SO2/SO3-convers ion is an indicator for the amount of SO3
pro-duced over the catalyst. SO3-concentration is measured up- and
downstre am of the catalysts in the ue gas according to the VDI
method 2462 [23] by controlled condensation of sulphuric acid
aerosols in a glass condenser at a temperat ure of 85 C. The tem-
perature is high enough to avoid condensation of water in uegas,
but low enough to condense SO3 as sulphuric acid aerosols.The
conversion is calculated considering the ue gas SO2-concen-tration,
which was measured by continuous monitors working onthe principle
of UV absorption, and the SO3-inlet- and outlet con- centration
accordin g to [22]:
SO2=SO3 conversionj23% cSO3out cSO3incSO2in 80 7
ring Journal 222 (2013) 274281The differe nt molar ratios of SO2
and SO3 accoun t for the factor of 80[%]. SO2/SO3-conversion
coefcient is calculate d similar to the DeNOx -activity.
-
3.3. Mercury oxidation
The concentration of elemental mercury is measured by
acontinuous mercury analyser (Hg2010, Semtech Metallurgy AB)working
with cold vapour atomic adsorption and a Zeeman background
correction. Oxidised mercury is determined as the difference
between total mercury, which is measured by wet chemical reduction
with tin(II)chloride solution upstream of the analyser, and the
concentration of elemental mercury. Mercury oxidation is therefore
calculated according to:
jHg0=Hg2 % cHg2out cHg2in
cHg0in 100 8
tioning is essential because there is a differenc e in activity
compar-
3000 mg/m 3 SO2 (not shown here) showed no further increase
inDeNOx-a ctivity.
The comparison of catalysts with different pitch at constant
area velocity leads to the conclusio n, that there is no inuence
ofthe catalysts geometry on DeNOx-acti vity evident, regardles sof
the sulphating of the active sites. An inuence of the thickness of
the wall cannot be seen from these measurements , which can be
referred to the fast reaction kinetics of the
NOx-reduction.According to [27], the reaction takes place only in
the rst 50 lmof the catalyst walls.
Fig. 4 shows the DeNOx-a ctivity measureme nt of all tested
catalyst determined at the bench reactor.
When comparing the results at the micro reactor and bench
reactor of the measure ments of the reference catalyst RC, it
isobvious that the DeNOx-a ctivities are approximately the
same.Deviation s can be explained by the higher NH3/NO-ratio in
micro
T. Schwmmle et al. / Chemical Engineering Journal 222 (2013)
274281 277ing fresh and preconditioned results. For commercial
application,sulphating is performed during the rst start-up of the
plant with installed new catalysts, so the test with ue gas SO2
shows the more realistic test conditions. Nevertheless, measure
ments atIt has to be stressed , that the in- and outlet total
mercury concen- tration has to be identica l, in order to ensure
steady state operation of the catalyst. These conditio ns are not
reached before sufcientprecond itioning time.
4. Results
4.1. DeNOx-activ ity
The most important reaction, which SCR-DeNOx -catalysts are
originally designed for, is the NOx-reduction. First of all, the
effect of the different pitch size on NOx-reduction was studied in
micro reactor at selected fresh catalyst samples. Measurem ents
were per- formed with the fresh catalyst sample in absence of
sulphur diox- ide in the ue gas. Following, the catalyst was
exposed to SO2 forsome days and the measure ments were repeated.
Fig. 3 showsthe results of these measurements .
The results of the experime nts in the presence and in the ab-
sence of sulphur dioxide differ signicantly. Fresh catalyst samples
show a limited sulphating of active sites. When exposed to ue gas
containing SO2, the Lewis and Brnsted acidity of the active sites
isincreased due to the adsorption of SO2 [24]. Being a Brnsted
base,NH3 is adsorbed at the acidic active sites, leading to an
increased DeNOx-acti vity. At TiO 2-based industrial SCR-catal
ysts, sulphating is only partially and reversibly and does not
cause any deactiva- tion, rather leading to enhanced DeNOx-a
ctivity [25]. In contrast to Svachula et al. [26], it can be stated
here, that for measuring the DeNOx-activity of commercial
SCR-DeNOx-ca talysts, precondi- Fig. 3. DeNOx-activity of catalysts
with different pitch at NH3/NO = 1.2 in micro reactor.reactor
measurement, leading to an excess of ammonia and slightly higher
DeNOx-acti vity. Catalyst reactors of full scale power plants are
designed with DeNOx-activity values of fresh honey- comb catalysts
of around 40 m/h [20], thus the determined values are quite
representat ive for standard commercial honeycomb catalysts.
These measureme nts also conrm the fast reaction kinetics
ofNOx-reduction, since there is no dependence on wall-thickness .In
contrast to the results of the micro reactor, a decrease of De-
NOx-acti vity was measure d with increasing pitch. This can be ex-
plained by differing ow conditions. The reaction process in the
catalyst is the result of external mass transfer (from bulk gas
phase to the surface of the catalyst), internal mass transfer
(diffusion ofreactants through internal pores to and from active
sites) and chemical reaction kinetics. Fig. 5 gives an insight in
ue gas owregimes through catalyst channels.
In the inlet-zon e, ue gas enters the channel, while the owprole
is not yet fully developed. There is mass exchange from the bulk
phase to the surface due to convection and diffusion.Downstr eam
the inlet zone, the laminar ow prole has been fully develope d. In
this prole, mass exchange is only due to diffusion,being limited
when compared to convection. In literature [20],the inlet-zone of a
catalyst channel is often called turbulent inlet zone. Therefore,
the rate of NOx-reduction is much greater in the inlet zone of the
catalyst. The dimensio ns of the test samples for micro reactor and
bench reactor result in different ratios of free width (d) to
length (l). For micro reactor catalysts, the d/l ratio equals to
0.03. In contrast to that, the d/l ratio of full scale mono- liths,
investigated with the bench reactor, varied between 0.005 and
0.008. Due to the short length of the micro reactor samples and the
relatively high area velocity, the fraction of the turbulentFig. 4.
DeNOx-activity of all 6 tested catalysts measured in bench reactor
at NH3/NO = 1.
-
It can be seen, that there is an increase in the
SO2/SO3-conver-sion for higher pitch in micro reactor measureme
nts. Unfortu- nately, the results gained in bench reactor measureme
nts, aspresente d in Fig. 6 show a less distinctive interrelation.
Further correlations have been performed and are shown in Figs. 7
and 8.
Unlike the characterist ics of the DeNOx-reac tion, a depende
ncy of the SO2/SO3-conversion on pitch cannot be traced back to the
clear width of the catalysts channels. Taking into account that
NOx-reduction is limited by external mass transfer and
diffusion,mass transfer controlled SO2-conversion would lead to
lower SO2/SO3-convers ion values. As shown in Fig. 7, there is a
linear depende ncy of the SO2/SO3-conversion on the catalysts wall
thick- ness. This effect can be explained by the relatively low
kinetics ofthis chemical reaction [32]. In contrast to
NOx-reduction, which proceeds rapidly when the reactants occupy the
active sites, the reaction of SO2 is compara tively slow. Thus, SO2
diffuses in the catalysts bulk and penetrates into the macroporou s
system ofthe catalyst wall. Thus, the reaction takes place within
the whole catalyst material. According to Dunn et al. [34], SO2
oxidation has
278 T. Schwmmle et al. / Chemical Engineering Journal 222 (2013)
274281zone is dominating, thus no dependence on pitch of DeNOx-
activity can be observed. In bench reactor (long samples), where
mass exchange is mainly controlled by diffusion, the impact of
owregime on DeNOx-activity is obvious, resulting in reduced
activity at larger channel openings. This effect of external
mass-transfer onNOx-reduction at honeycomb catalysts was also
observed and described by [26,28,2931]. They considered not only
the inter- phase mass transfer, but also the intra-phase mass
transfer in the catalyst porous system. Inter-phase mass transfer
can be described by correlations of the Sherwood number and is
described in detail in the given references.
4.2. SO2/SO3-conversion
SO2/SO3-conversion was determined by micro reactor and bench
reactor measureme nts in the absence of ammonia in uegas after at
least 48 h of preconditioning. Hence, the values presented in Fig.
6 show the maximum values at the measured SO2-concentrat ion.
For decreasing ue gas SO2-concentrat ions, the relative
SO2/SO3-conversion will increase, however the absolute
SO3-concentratio n downstream of the SCR is decreasing. This is in
good agreement with previous ndings in [32,33]. Differences between
the micro reactor and bench reactor values can be referred
totemperature deviations. The effect of temperat ure on
SO2/SO3-conversion was discussed in a previous work [33], pointing
out the exponential dependence. After nishing the test series in
the
Fig. 5. Flow conditions in a catalyst channel.micro reactor, a
deviation in temperature of 14 C above 380 Cduring the
SO3-measurement es was determined. This explains the deviations
between the measureme nts in micro- and bench reactor.
Fig. 6. SO2/SO3-conversion coefcient determined in micro- and
bench reactor without NH3, NO.Fig. 7. Correlations of
SO2/SO3-conversion measured in micro reactor.Fig. 8. Correlation of
SO2/SO3-conversion measured in bench reactor.
-
a very low turnover frequency, caused by the inefcient adsorptio
nof SO2 on the acidic surface vanadia species under reaction condi-
tions. Correlations with limited amount of variations have also
been performed by Svachula et al. [32], who concluded, that there
is a linear dependency of SO2 oxidation on catalyst wall
thickness.Since the catalysts wall and therefore the whole volume
of the cat- alyst is responsible for the increase in
SO2/SO3-convers ion, a corre- lation of the SO2/SO3-conversion over
the weight per channel has been performed in Fig. 7. The weight per
channel ratio was calcu- lated by dividing the total mass of the
sample by its amount ofchannels and provides plausible results,
tracing back SO2/SO3-con-version directly to the catalysts mass.
For bench-reactor SO2/SO3-conversion measureme nts, a similar
correlation with catalysts
lyst, due to fewer active sites occupied by the DeNOx-reac tion
because of superior mass exchange . Different values of mercury
oxidation in Seniors study compared to the values presented here
can be referred to lower residence time in the catalyst at the
exper- iments, leading to lower mercury oxidation at the
samples.
4.4. Combined reactions
In addition to the research on single reactions at the
catalyst,which were performed in order to study the reactions in
detail,experime nts observing the interaction of the reactants
involved were conducted. Fig. 11 shows the SO2/SO3-convers ion
during par- allel NOx-reduction measured at catalyst P8.2 in the
micro reactor.
There is a linear increase of NOx-reduction for increasing
ammonia /nitrogen oxide ratio, proving that the availabili ty
ofammonia is the limiting factor of the reaction up to the
stoichiom- etric ratio of 1. At this point, reaction order with
regard to NH3 be-comes zero. NOx-reduction is not exactly
proportional to NH3/NO-
T. Schwmmle et al. / Chemical Engineemass was performed (Fig.
8).For a comparative presentation , the SO2/SO3-convers ion was
di-
vided by the catalysts sample mass. These mass-specic values
show very similar results for all catalysts investiga ted.
However,the value of catalyst RC is quite high, which can possibly
be traced back to a measure ment error. All these ndings indicate
that SO2/SO3-conversion is directly dependent on catalysts mass.
However ,referring SO2/SO3-conversion to catalysts surface, as
performed atthe conversion coefcient in Fig. 6, is expected to lead
to an in- crease with increasing wall thickness. As a further
consequence,comparison of non-mass-s pecic values of
SO2/SO3-conversionfor catalyst materials with different chemical
compositi on can only be performed for identical catalyst wall
thickness.
4.3. Mercury oxidation
As the third reaction, mercury oxidation at the catalysts with
different pitch was studied. In Fig. 9, mercury oxidation at the
com- mercial reference catalyst at micro reactor test conditions
(withoutNH3, NO and SO2) is related to ue gas HCl-concen
tration.
Despite mercury ue gas concentratio n is tremend ously lower
compared to the ue gas HCl-concen tration (cHCl/cHg = 1000), there
is a dependency on the HCl-content. At HCl concentratio ns higher
than 50 mg/m 3, the increase in mercury oxidation reaches alimiting
value. Therefore, the concentratio n of 100 mg/m 3 was cho- sen as
the upper limit of the experiments. Thus, the studied range
ofHCl-concen tration covers almost the whole range of full scale
power plant HCl-concen trations, which is strongly depended oncoal
origin. Kolker et al. [35] presented a range of 105000 mg Cl/ kg
coal, resulting in ue gas concentratio ns of about 1500mg HCl/m 3.
However, most of the coals red in boilers globally have low
Cl-content, resulting in ue gas HCl-concen trations lower than 100
mg/m 3. It is commonly accepted, that HCl plays an important role
in mercury oxidation [13,14,36], but no evidence was given,why
there is such an abundan ce of HCl needed, even if there are no
other possibly disturbing ue gas components involved.Fig. 9. Effect
of ue gas HCl-concentration on mercury oxidation.Fig. 10 shows the
correlation of the mercury oxidation with cat- alysts wall
thickness at different HCl-concentra tions.
For all investigated HCl-concen trations, there is a slight
increase of mercury oxidation with catalysts wall thickness and
pitch. Asshown at the DeNOx-reactio n, a fully diffusion controlle
d reaction results in decreasing or at least constant values for
increasing pitch/wal l thickness at the micro reactor. The results
reveal that mercury oxidation does not only take place at the
surface of the catalyst, indicating slow chemical reaction
kinetics. Thus, the reac- tants (mercury and/or HCl) are able to
penetrate into the catalysts macropo rous system. These results are
in contrast to the theoreti- cal calculations performed by Niksa
and Fujiwara [13], showing in- creased mercury oxidation at lower
pitch sizes respectivel y low channel openings . This was explained
by mass transport phenom- ena. Senior [37] developed a model for
mercury oxidation at SCR- DeNOx-c atalysts and concluded, based on
data of [9], that there islower mercury oxidation at high dust
catalysts with greater pitch compare d to low dust catalysts with
lower pitch. However, it has to be kept in mind that in [37], only
two catalysts with a difference in wall thickness of 0.4 mm were
considered. In this study, more catalysts with a range of over 0.6
mm were measured. Addition ally,Seniors calculations have been
performed at a NH3/NO-ratio of 1.Taking into account the results
gained at the DeNOx-activity mea- suremen ts above and the
assumption, that DeNOx-activity and mercury oxidation takes place
at the same active sites, it is obvi- ous, that there is higher
mercury oxidation at the low pitch cata-
Fig. 10. Effect of catalysts wall thickness on mercury oxidation
of SCR-DeNOx catalyst.
ring Journal 222 (2013) 274281 279ratio, which can also be
referred to the low residence time of the gas inside the catalyst.
DeNOx-acti vity was also measured with ue gas mercury and HCl (100
mg/m 3). The results showed no
-
adsorbed ammonia constrain s the diffusion of SO2 through the
cat-
the main relevant reactions at the SCR-DeNOx-ca talyst, but
for
280 T. Schwmmle et al. / Chemical Engineealysts walls.
Measurements and calculatio ns by Orsenigo et al. [24]showed the
same behaviour, but the inhibition occurs rstlyaround NH3/NO = 1.0
at an area velocity of 10 m/h with a slight tendency to lower
ratios at increased area velocities. Due to the high area
velocities (AV = 27.5 m/h) in this study, the inhibition occurs at
relatively low NH3/NO ratios because of the high NH3-coverage of
the active sites. The interaction of DeNOx- reaction and
SO2/SO3-conversion will also take place in full scale SCR-reactor s
of coal red power plants. In the rst catalysts layer,correspondi ng
to high NH3/NO ratios, the SO2/SO3-conversion islow and increases
signicantly at the last catalyst layer where noammonia is present.
Thus, the values of SO2/SO3-conversion inFig. 6 represent
conditions in the last catalyst layer of the full scale high-dust
SCR-DeNOx reactor.
4.5. Pressure drop
In Fig. 12, pressure drop over the tested catalysts measured
inbench reactor and normalised to 1000 mm length is shown.
Pressure drop increases with decreasing clear width respec-
tively cell opening. This will directly inuence the operating costs
of the power plant by increased power consumptio n of the induced
inuence of the reactants of the mercury oxidation reaction
onDeNOx-acti vity, which can be explained by the low mercury con-
centration and the acidic nature of the active sites.
However,SO2/SO3-conversion is inhibited by parallel DeNOx-reactio
n. For the NH3/NO-ratio of 0.8, the rate of SO3 formation is almost
zero.Active sites are preferent ially occupied by ammonia, hence
SO2cannot be converte d to SO3. Moreover, it can be assumed,
that
Fig. 11. SO2/SO3-conversion at parallel NOx-reduction.draught.
Due to combined variation of pitch and wall thickness, the void
fraction, which is also relevant for pressure drop, stays
almost
paramete rs of honeycomb catalysts has been performed .
Although
Fig. 12. Pressure drop over tested catalysts measured at
bench-reactor and normalised to 1000 mm length.there are commonl y
accepted correlations for the comparison ofdifferent test
conditions based on reaction kinetics (e.g. DeNOx- the operation of
SCR units of full scale power plants, there are sev- eral other
factors to be considered. Due to the fact, that in this study a
combined variation of pitch and wall thickness has been performed ,
the effect of the two factors has to be discussed separately .
SCR-DeNO x-catalysts are -regarding the material usage- very
ineffective related to DeNOx-perfor mance. Only a very thin surface
layer of around 50 lm is needed for the reaction to take place.
Nev- ertheless, walls of commerc ial SCR-DeNOx -catalysts have a
thick- ness of around 1 mm. As a result, the monolith wall
thickness could be reduced signicantly without affecting the
DeNOx-perfor -mance. Furthermor e, it would be benecial to reduce
the wall thickness in order to reduce the undesired SO2/SO3
conversionand catalyst cost. The pressure drop will also be lower
due to high- er void fraction. Unfortunate ly, mercury oxidation at
the catalysts with lower wall thicknesses would be reduced
simultaneou sly. Onthe other hand, regarding the pitch of the
catalyst, it would bedesirable for increased NOx-reduction to apply
small pitch cata- lysts due to shorter distances of diffusion and
gassolid mass transfer limitations of this fast reaction.
However , there are several constrain ts from practical point
ofview, mainly related to applicati on limitations and the catalyst
production process. The applicable channel clear width (or pitch)is
mainly determined by the particulate content of the ue gas,the
characterist ics of the dust, and the allowable pressure drop
across the SCR reactor. The more dust, the larger the pitch size
inorder to avoid or minimize dust deposition and catalyst plugging
.Catalyst channels in high-dust applicati ons (where the catalyst
islocated immediatel y after the boiler, processing the full dust
load- ing leaving the boiler) will therefore be signicantly larger
than ina tail-end installati on (the catalyst is located downstre
am of a par- ticulate collection device and a ue gas desulphuriz
ation FGD sys- tem). The honeycomb catalyst wall thickness for
high-dust applicati ons at coal-red boilers shall be within a
certain range (typically between 0.8 and 1.1 mm) to maintain
mechanical resis- tance in operation and during a possible catalyst
regeneration pro- cess. For tail-end or gas red applicati ons small
pitch catalysts (24 mm) with a reduced wall thickness (0.250.5 mm)
can be used.
Honeyco mb catalysts are obtained by extruding a ceramic paste
made by catalytic material followed by a delicate drying
step.Therefore a certain ratio between inner wall thickness and
channel clear width or respectively a certain range of the monolith
void fraction (typically between 70% and 80%) shall be maintain ed
for successfu l production. Also the (transversal) compression
strength of the nal honeycomb catalyst is strongly affected by the
number of cells and the inner wall thickness. Sufcient compress ion
strength is required for packing of the catalyst monoliths into the
steel frames.
6. Conclusion
A systematical study on the effect of the variation of geometri
cal constant . Larger channels at P10 lead to lower pressure drop
com- pared to P6.
5. Discussion
There is a great inuence of honeycomb catalysts geometry on
ring Journal 222 (2013) 274281activity) available, the specic
test condition s have to be consid- ered. Micro reactor studies in
laboratory with small samples are asuitable tool for systemati c
studies, especially regarding still not
-
fully claried reaction mechanism s and interactions of
mercury
reactants use the whole catalysts bulk and not only the
surface
reduced and the wall thickness increased to a certain extent.
How-
[7] J.P. Chen, R.T. Yang, Role of WO3 in mixed V2O5WO3/TiO2
catalysts for
T. Schwmmle et al. / Chemical Engineering Journal 222 (2013)
274281 281ever, this would increase the material costs due to
higher amount of catalyst material needed and would also lead to
higher pressure drop and a higher risk of catalyst blocking by y
ash. Unfortu- nately, this would also boost SO2/SO3-conversion at
the catalysts,which is highly undesired by power plant operators.
Therefore,increasing of mercury oxidation by variation of catalysts
pitch orwall thickness is not possible without drawbacks. However,
further investigatio ns on the interaction of the reactants with
the catalysts material will be performed in order to gain a greater
insight into the reaction mechanis m and its dependencies. Still,
there are methods available for selectively increasing mercury
oxidation byvariation of (chemical) composition of the catalyst,
which are cur- rently under research in the ongoing project.
Acknowled gements
The authors greatly thank IBIDEN Porzellanfabrik Frauenthal GmbH
for providing the catalyst samples and E.ON New Build and Technolo
gy GmbH for performing the bench-react or tests.The work was funded
by the European Commission within the DEVCAT project under the
Research Fund for Coal and Steel ofthe European Commission
(RFCR-CT-2010-00012).
References
[1] IEA Statistics: Electricity Information 2012, International
Energy Agency, 2012.[2] M. Schrod, J. Semel, R. Steiner, Verfahren
zur Minderung von NOx-Emissionen
in Rauchgasen, Chem. Ing. Tech. 57 (9) (1985) 717727.[3] H.
Bosch, F. Janssen, Catal. Today 2 (1988) 369531.[4] P. Forzatti,
Present status and perspectives in de-NOx SCR catalysts, Appl.
Catal. A: General 222 (2001) 221236.[5] I. Nova, A. Beretta, G.
Groppi, L. Lietti, E. Tronconi, P. Forzatti, Monolithic
Catalysts for NOx Removal from Stationary Sources, in: A.
Cybolski, J. Moulijn (Eds.), Structured Catalysts and Reactors, CRC
Taylor & Francis, 2006, pp. 171210.
[6] L. Lietti, P. Forzatti, F. Bregani, Steady-state and
transient reactivity study ofTiO2-supported V2O5WO3 De-NOx
catalysts: relevance of the vanadiumtungsten interaction on the
catalytic activity, Ind. Eng. Chem. Res. 35 (1996)38843892.(unlike
DeNOx reaction) for their reaction.SO2/SO3-conversion is
signicantly inhibited by ue gas ammo-
nia, which is adsorbed at the catalyst. Hence, the DeNOx-reactio
ntakes place at the rst catalysts layers of the full scale catalyst
reac- tor. Signicant conversion of ue gas SO2 to SO3 will only
occur inthe last catalyst layer where the ammonia concentratio n is
almost zero.
To enhance the desired reactions of DeNOx and mercury oxida-
tion, the catalysts pitch of commercial catalysts could be slightly
oxidation. Bench reactor test of full scale monoliths are a commer-
cial tool for suppliers and catalysts manufactur ers to get
reliable,near full scale comparis on data in order to screen
guarantee values.However, considering the results of micro reactor
and bench reactor tests presented, the ow regime and the mass
transfer has to beconsidered.
DeNOx-acti vity decrease s when catalysts clear width respec-
tively the channel diameter increases. This effect could be traced
back to mass transfer limitations. Increased wall thickness of
cata- lysts could lead to higher mercury oxidation, but the
drawback is atremendous ly higher SO2/SO3-conversion which almost
linearly increases with wall thickness in the studied range. Both
reactions are to a certain extent limited by chemical reaction
kinetics. The selective catalytic reduction of nitric oxide with
ammonia, Appl. Catal. A:General 80 (1992) 135148.
[8] G. Scheffknecht, S. Farr, B. Heidel, T. Schwmmle, K.
Brechtel, Quecksilber aus thermischen Kraftwerken: Freisetzungs-
und Umwandlungsmechanismen sowie Mglichkeiten zur Minderung, Chem.
Ing. Tech. 84 (2012) 10411051.
[9] M. Hocquel, The behaviour and fate of mercury in coal-red
power plants with downstream air pollution control devices,
VDI-Verlag, Dsseldorf, 2004.
[10] C. Martel, Brennstoff- und lastspezische Untersuchungen von
Schwermetallen in Kohlestaubfeuerung, VDI-Verlag, Dsseldorf,
2000.
[11] H. Gutberlet, A. Spiesberger, F. Kastner, J. Tembrink, Zum
Verhalten des Spurenelementes Quecksilber in Steinkohlefeuerungen
mit Rauchgasreinigungsanlagen, VGB Kraftwerkstechnik 72 (1992)
636641.
[12] D. Laudal, Effect of SCR on Mercury, 2002 Field Studies
Update, UNDEERC, Final Report, 2003.
[13] S. Niksa, N. Fujiwara, A predictive mechanism for mercury
oxidation onselective catalytic reduction catalysts under
coal-derived ue gas, J. Air Waste Manage. Assoc. 55 (2005)
18661875.
[14] H.J. Hong, S.W. Ham, M.H. Kim, et al., Characteristics of
commercial selective catalytic reduction catalyst for the oxidation
of gaseous elemental mercury with respect to reaction conditions,
Korean J. Chem. Eng. (2010) 16.
[15] B.W. Vosteen, R. Kanefke, H. Kser, Bromine-enhanced mercury
abatement from combustion ue gases-recent industrial applications
and laboratory research, VGB PowerTech 86 (2006) 7075.
[16] B.J. Aylett, Comprehensive Inorganic Chemistry, Pergamon
Press, Oxford, 1975.[17] J. Fahlke, A. Bursik, Impact of the
state-of-the-art of ue gas cleaning on
mercury species emissions from coal-red steam generators, Water,
Air Soil Pollut. 80 (1995) 209215.
[18] UNEP: Process Optimisation Guidance for Reducing Mercury
Emissions from Coal Combustion, Geneva, Switzerland, 2010.
[19] B. Basu, J. Butler, M. Leon, Biomass co-ring options on the
emission reduction and electricity generation costs in coal-red
power plants, Renew. Energ. 36(2011) 282288.
[20] VGB Powertech: VGB Standard Bauart, Betrieb und Wartung von
Rauchgasentstickungsanlagen (DeNOx),
VGB-Standard-S-014-2011-DE,Essen, 2011.
[21] I. Nova, L. Lietti, E. Tronconi, P. Forzatti, Dynamics of
SCR reaction over a TiO 2-supported vanacia tungsta commercial
catalyst, Catal. Today 60 (2000) 7382.
[22] VGB guideline VGB-R 302 H e: guideline for the testing of
DeNOx catalysts.Second ed., VGB, Essen, 1998.
[23] VDI Norm 2462 Bl. 2: Messen gasfrmiger Emissionen,
Bestimmung von Schwefeloxiden. VDI/DINHandbuch Reinhaltung der
Luft, Band 5, Beuth Verlag, Berlin, 2011.
[24] C. Orsenigo, A. Beretta, P. Forzatti, et al., Theoretical
and experimental study ofthe interaction between NOx reduction and
SO2 oxidation over DeNOx-SCR catalysts, Catal. Today 27 (1996)
1521.
[25] G. Busca, L. Lietti, G. Ramis, F. Berti, Chemical and
mechanistic aspects of the selective catalytic reduction of NOx by
ammonia over oxide catalysts: a review,Appl. Catal. B: Environ. 18
(1998) 136.
[26] J. Svachula, N. Ferlazzo, P. Forzatti, E. Tronconi,
Selective reduction of NOx byNH3 over honeycomb DeNOxing catalysts,
Ind. Eng. Chem. Res. 32 (1993)10531060.
[27] J. Beck, The behaviour of phosphorous compounds and their
effect an De-NOx- catalysts in coal red power plants, VDI-Verlag,
Dsseldorf, 2007.
[28] I. Binder-Begsteiger, G. Herzog, E. Megla, M. Tomann-Rosos,
Zur Kinetik der Denox-Reaktion an TiO 2/WO3-Wabenkatalysatoren,
Chem. Ing. Tech. 62(1990) 6061.
[29] I. Binder-Begsteiger, Improved emission control due to a
new generation ofhigh-void-fraction SCR-DeNOx catalysts, Catal.
Today 27 (1996) 38.
[30] E. Tronconi, A. Beretta, The role of inter- and intra-phase
mass transfer in the SCR-DeNOx-reaction over catalysts of different
shapes, Catal. Today 52 (1992)249258.
[31] J. Beckmann, L. Hegedus, Design of monolith catalysts or
power plant NOxemission control, Ind. Eng. Chem. Res. 30 (1991)
969978.
[32] J. Svachula, L.J. Alemany, N. Ferlazzo, et al., Oxidation
of SO2SO3 overhoneycomb DeNoxing catalysts, Ind. Eng. Chem. Res. 32
(1993) 826834.
[33] T. Schwaemmle, B. Heidel, K. Brechtel, G. Scheffknecht,
Study of the effect ofnewly developed mercury oxidation catalysts
on the DeNOx-activity and SO2SO3-conversion, Fuel 101 (2012)
179186.
[34] J. Dunn, H.G. Stenger Jr., I.E. Wachs, Oxidation of sulfur
dioxide over supported vanadia catalysts: molecular
structurereactivity relationships and reaction kinetics, Catal.
Today 51 (1999) 301318.
[35] A. Kolker, C.L. Senior, J.C. Quick, Mercury in coal and the
impact of coal quality on mercury emissions from combustion
systems, Appl. Geochem. 21 (2006)18211836.
[36] H. Kamata, S. Ueno, N. Sato, T. Naito, Mercury oxidation by
hydrochloric acid over TiO 2 supported metal oxide catalysts in
coal combustion ue gas, Fuel Process. Technol. 90 (2009)
947951.
[37] C.L. Senior, Oxidation of mercury across SCR catalysts in
coal-red power plants, J. Air Waste Manage. Assoc. 56 (2006)
2331.
Influence of geometrical parameters of honeycomb commercial
SCR-DeNOx-catalysts on DeNOx-activity, mercury oxidation and
SO2/SO3-conversion1 Introduction2 Experimental method2.1
Catalysts2.2 Test setup2.2.1 Micro reactor2.2.2 Bench reactor
3 Determined parameters3.1 DeNOx-activity3.2
SO2/SO3-conversion3.3 Mercury oxidation
4 Results4.1 DeNOx-activity4.2 SO2/SO3-conversion4.3 Mercury
oxidation4.4 Combined reactions4.5 Pressure drop
5 Discussion6 ConclusionAcknowledgementsReferences