Chemical Engineering Science 66 (2011) 5192–5203
`
`Contents lists available at ScienceDirect
`
`Chemical Engineering Science
`
`journal homepage: www.elsevier.com/locate/ces
`
`Experimental study of mass transfer limitations in Fe- and Cu-zeolite-based
`NH3-SCR monolithic catalysts
`
`Pranit S. Metkar, Vemuri Balakotaiah n, Michael P. Harold nn
`
`Department of Chemical and Biomolecular Engineering, University of Houston, 4800 Calhoun Road, Houston, TX 77204, USA
`
`a r t i c l e i n f o
`
`a b s t r a c t
`
`Article history:
`Received 24 May 2011
`Received in revised form
`30 June 2011
`Accepted 6 July 2011
`Available online 19 July 2011
`
`Keywords:
`Catalysis
`Chemical reactors
`Mass transfer
`Reaction engineering
`Washcoat diffusion
`External mass transfer
`
`An experimental study of steady-state selective catalytic reduction (SCR) of NOx with NH3 on both
`Fe-ZSM-5 and Cu-ZSM-5 monolithic catalysts was carried out to investigate the extent of mass transfer
`limitations in various SCR reactions. Catalysts with different washcoat loadings, washcoat thicknesses
`and lengths were synthesized for this purpose. SCR system reactions examined included NO oxidation,
`NH3 oxidation, standard SCR, fast SCR and NO2 SCR. Comparisons of conversions obtained on catalysts
`with the same washcoat volumes but different washcoat thicknesses indicated the presence of
`washcoat diffusion limitations. NH3 oxidation, an important side reaction in SCR system, showed the
`presence of washcoat diffusion limitations starting at 350 1C on Fe-zeolite and 300 1C on Cu-zeolite
`catalysts. Washcoat diffusion limitations were observed for the standard SCR reaction (NH3þNOþO2)
`on both Fe-zeolite (Z350 1C) and Cu-zeolite (Z250 1C). For the fast (NH3þNOþNO2) and NO2 SCR
`(NH3þNO2) reactions, diffusion limitations were observed throughout the temperature range explored
`(200–550 1C). The experimental findings are corroborated by theoretical analyses. Even though the
`experimentally observed differences in conversions clearly indicate the presence of washcoat diffusion
`limitations, the contribution of external mass transfer was also found to be important under certain
`conditions. The transition temperatures for shifts in controlling regimes from kinetic to washcoat
`diffusion to external mass transfer are determined using simplified kinetics. The findings indicate the
`necessity of inclusion of mass transfer limitations in SCR modeling, catalyst design and optimization.
`& 2011 Elsevier Ltd. All rights reserved.
`
`1.
`
`Introduction
`
`Increasing transportation fuel prices have increased the demand
`for diesel powered vehicles, which are more fuel efficient than their
`gasoline counterparts. Diesel exhaust contains volatile hydrocarbons,
`particulate matter (PM) and NOx, mostly in the form of NO. In
`response to increasingly stringent EPA standards for NOx, a host of
`NOx reduction technologies are being developed. Selective catalytic
`reduction (SCR) of NOx with NH3 generated from onboard thermal
`decomposition of urea has gained considerable attention in the last
`few years and has proven to be an effective catalytic process for NOx
`reduction. Various catalysts have been studied and researched for the
`ammonia-based SCR technique. Vanadia-based catalysts are widely
`investigated and used for this purpose (Busca et al., 1998; Ciardelli
`et al., 2007; Madia et al., 2002a). However, this catalyst has been
`found to have reduced stability at higher temperatures and the
`toxicity of vanadia is problematic. For these reasons, recent research
`has focused on Fe- and Cu-exchanged zeolite catalysts, both of which
`are found to have a high NOx removal efficiency over a wide
`
`n Corresponding author. Tel.: þ1 713 743 4318.
`nn Corresponding author. Tel.: þ1 713 743 4322.
`E-mail addresses: bala@uh.edu (V. Balakotaiah),
`mharold@uh.edu (M.P. Harold).
`
`0009-2509/$ - see front matter & 2011 Elsevier Ltd. All rights reserved.
`doi:10.1016/j.ces.2011.07.014
`
`temperature range. The performance of Cu- and Fe-zeolite catalysts
`has been reported in literature studies (Delahay et al., 1997; Devadas
`et al., 2006; Grossale et al., 2008; Long and Yang, 2002; Metkar et al.,
`2011; Schwidder et al., 2008; Sjovall et al., 2006; Sun et al., 2001). In
`general, Cu-based catalysts have higher NOx reduction activity at
`lower temperatures (o350 1C) whereas Fe-based catalysts are more
`active at higher temperatures (43501C). The overall chemistry of
`various SCR reactions is well established although predictive kinetic
`and reactor models still need to be developed. Moreover, the extent
`of mass transfer limitations has yet to be resolved. We return to this
`point below after highlighting the main features of the chemistry.
`
`The selective catalytic reduction involves the following set of
`reactions which are divided into three main categories:
`
` Standard SCR reaction: This reaction involves NO and NH3
`reacting in the presence of O2:
`4NH3þ4NOþO2-4N2þ6H2O, DH¼–4.07  105 J/mol NH3 (1)
` Fast SCR reaction: When both NO and NO2 in the feed react
`simultaneously to produce N2 and H2O; it is called the ‘‘fast
`SCR’’ reaction (2) because (in the practical range of tempera-
`tures) it is much faster than the standard SCR reaction (1):
`2NH3þNOþNO2-2N2þ3H2O, DH¼–3.78  105 J/mol NH3 (2)
`
`BASF-2031.001
`
`

`
`P.S. Metkar et al. / Chemical Engineering Science 66 (2011) 5192–5203
`
`5193
`
` NO2 SCR reaction: This involves reaction between NO2 and NH3
`
`and is given by
`4NH3þ3NO2-3.5N2þ6H2O, DH¼–3.41  105 J/mol NH3
`
`(3)
`
`Along with these main reactions, additional side reactions occur.
`NH3 oxidation is a key side reaction as it competes with the SCR
`reaction for the reductant ammonia and hence it is considered as an
`undesired side reaction. On Fe- and Cu-zeolite catalysts, most of the
`NH3 is selectively oxidized to N2 by
`4NH3þ3O2-2N2þ6H2O, DH¼–3.12  105 J/mol NH3
`
`(4)
`
`The oxidation of NO to NO2 occurs in the temperature range of
`interest (T4150 1C) and is desirable because NO2 is more effec-
`tively reduced by NH3 than is NO:
`NOþ½O2’-NO2, DH¼–5.7  104 J/mol NO
`
`(5)
`
`On the other hand, the existence of NO2 complicates the reaction
`system. In particular, ammonium nitrate formation occurs:
`2NH3þ2NO2-N2þNH4NO3þH2O,DH¼–2.91  105 J/mol NH3 (6)
`
`A mechanism for the ammonium nitrate formation in the pre-
`sence of NO2 has been reported on vanadia- and various zeolite-
`based catalysts (Grossale et al., 2008; Koebel et al., 2001; Madia
`et al., 2002b). Ammonium nitrate is decomposed to N2O at higher
`temperatures (Z200 1C):
`NH4NO3-N2Oþ2H2O, DH¼–3.66  104 J/mol NH4NO3
`
`(7)
`
`In the past several years, many literature studies have focused
`on the performance of the emerging Cu- and Fe-based zeolite
`catalysts. The data from these studies have been used in the
`development of kinetic models for SCR reactions (Chatterjee et al.,
`2007; Olsson et al., 2008). However, in order to develop an
`intrinsic kinetic model, it is prudent to measure rates in a pure
`kinetic regime so that the impacts of external and internal mass
`transfer limitations on the observed kinetics are negligible. By
`now, the importance of mass transfer processes occurring in
`monolithic reactions is well understood (Balakotaiah and West,
`2002; West et al., 2003). Washcoat diffusion limitations (known
`more generally as internal mass transfer limitations) are known
`to play a key role and can be rate limiting in various washcoated
`monolithic catalytic reactions (Joshi et al., 2010; Joshi et al., 2011).
`However, washcoat diffusion along with external mass transfer have
`been either ignored or ruled out in the development of the existing
`kinetic models for SCR catalysts.
`Indeed, few SCR studies are
`available, which describe the existence of diffusion limitations. In
`our recent study (Metkar et al., 2011), we showed the existence of
`washcoat diffusion limitations for the standard SCR reaction cata-
`lyzed by Fe-zeolite at higher temperatures (Z350 1C). Olsson et al.
`(2008) studied the standard SCR reaction on Cu-zeolite and reported
`NO conversion of more than 95% at temperatures of as low as
`200 1C. They used catalysts with different washcoat loadings to
`investigate the presence of washcoat diffusion limitations. However,
`with very high conversions (100%) of NO at temperatures above
`300 1C, the presence of diffusion limitations was inconclusive and
`thus diffusional gradients were not included in their kinetic model.
`A recent study by Nova et al. (2011) indicated the presence of mass
`transfer limitations for Cu-zeolite catalysts. They used monolith
`catalysts having different cell densities (200, 400 and 600 CPSI) but
`the same washcoat loadings. The NH3 conversion was found to have
`a substantial difference between the 200 CPSI and 400 CPSI catalysts.
`The difference was negligible when comparing the 400 and 600 CPSI
`catalysts. The authors attributed the differences to mass transfer
`limitations for the SCR reaction on Cu-zeolite catalysts. Chatterjee
`
`et al. (2007) found comparable NOx reduction activity on crushed
`monolith (zeolite) powder and washcoated zeolite catalysts while
`studying the standard SCR reaction. Thus they neglected the inter-
`phase and intraphase diffusion limitations in their model. Collec-
`tively, these studies convey some uncertainty of the extent of mass
`transfer limitations in the SCR reaction system.
`Therefore, the objective of the current study is to investigate
`the extent of internal and external mass transfer limitations
`during several representative reactions occurring in the SCR
`reaction systems. A systematic study of various SCR reactions
`was carried out on both Fe-ZSM-5 and Cu-ZSM-5 washcoated
`catalysts. For this purpose, catalysts with various washcoat loadings,
`washcoat thicknesses and lengths were synthesized in our labora-
`tory. The reactions studied included: NO oxidation, ammonia
`oxidation, standard SCR (NH3þNO), fast SCR (NH3þNOþNO2) and
`NO2 SCR. For the NO2 SCR reaction, the effect of feed water and O2
`was considered separately. A detailed analyses of characteristic
`times for various transport processes and of the Weisz–Prater
`modulus are presented to support the experimental findings. The
`effect of temperature and effective washcoat diffusivities on transi-
`tions between various regimes (from kinetic to washcoat diffusion
`to external mass transfer) is presented using simplified kinetics and
`related to experimental observations.
`
`2. Experimental description
`
`2.1. Catalyst preparation
`
`The SCR of NOx by NH3 reaction system was studied on Fe-
`ZSM-5 and Cu-ZSM-5 washcoated monolith catalysts synthesized
`in-house. The Cu-ZSM-5 catalyst was prepared by conventional
`þ
`form of zeolite (NH4-ZSM-5,
`ion-exchange starting with the NH4
`Sud-Chemie Munich, Germany) powder having a Si/Al ratio of 25.
`þ
`þ
`form was converted to protonated (H
`) form by cal-
`The NH4
`cination (500 1C for 5 h). The H-ZSM-5 powder was then con-
`verted to Na-ZSM-5 by ion-exchange by continuously stirring it in
`a 0.1 M NaNO3 solution, followed by several steps of filtration and
`drying. In the final step, Na-ZSM-5 was ion-exchanged with
`0.02 M copper acetate solution to get Cu-ZSM-5. The Cu-ZSM-5
`powder thus obtained was calcined for 5 h at 500 1C. The method
`used for ion-exchange was similar to described in the literature
`(Sjovall et al., 2006). Fe-ZSM-5 powder, provided by Sud-Chemie
`(Munich, Germany), was used to synthesize Fe-based washcoated
`catalysts. The washcoated catalysts were prepared by dipping a
`blank cordierite monolith in slurry comprising a mixture of either
`Fe-ZSM-5 powder or Cu-ZSM-5 powder and g-alumina particles.
`More detail about the washcoating procedure can be found in our
`earlier study (Metkar et al., 2011). Characterization included
`inductive coupled plasma (ICP) and scanning electron microscopy
`(SEM) to provide information about the elemental composition
`and washcoat thickness, respectively.
`
`2.2. Bench-scale reactor setup
`
`The experimental setup was the same one used in our previous
`studies (Clayton et al., 2008; Kabin et al., 2004). It included a gas
`supply system, a reactor system, an analytical system and a data
`acquisition system. The monolith catalysts were wrapped with
`ceramic paper and inserted inside a quartz tube reactor (40.6 cm
`long, 0.81 cm inner diameter, 1.27 cm outer diameter) which was
`mounted in a tube furnace coupled to a temperature controller.
`An FT-IR spectrometer (Thermo-Nicolet, Nexus 470) and a quad-
`rupole mass spectrometer (QMS; MKS Spectra Products; Cirrus
`LM99) were positioned downstream of the reactor to analyze the
`effluent gases.
`
`BASF-2031.002
`
`

`
`5194
`
`P.S. Metkar et al. / Chemical Engineering Science 66 (2011) 5192–5203
`
`Monolith catalysts with different dimensions (lengths, wash-
`coat loadings, number of channels) were used to study the
`washcoat diffusion limitations for various reactions. Before start-
`ing each experiment, each catalyst was pretreated by flowing 5%
`O2 in Ar with total flow rate of 1000 sccm while keeping the
`catalyst temperature constant at 500 1C for 30 min. The catalyst
`temperature was then brought back to the room temperature
`before the start of each experiment.
`In order to study the diffusion limitations in various SCR
`reactions, several Fe- and Cu-zeolite catalyst samples were
`synthesized with prescribed washcoat loadings. These catalysts
`are named as FeZ-XX and CuZ-XX where -XX denotes the wash-
`coat loading (weight%). Most experiments were carried out in the
`temperature range of 150–550 1C. We waited about 30 min at
`each temperature until steady state was reached.
`For the NH3 oxidation reaction, the catalysts had 28 channels
`and the total flow rate was kept constant at 1000 sccm. For the
`purpose of keeping the same space velocity per unit mass of
`catalyst (i.e. keeping the W/F, mass of catalyst/molar flow rate,
`ratio constant), 2 cm length of FeZ-11 and 1 cm length of FeZ-22
`catalyst were used. A similar approach was used for this study on
`Cu-zeolite catalysts, where CuZ-10 (2 cm) and CuZ-20 (1 cm)
`samples were used. The feed consisted of 500 ppm NH3, 5% O2
`and 2% water.
`For the NO oxidation reaction, same Fe-zeolite catalysts (2 cm
`long FeZ-11 and 1 cm long FeZ-22) were used. The feed consisted
`of 500 ppm NO, 5% O2. For the standard SCR, using the same
`catalysts the feed consisted of 500 ppm NO, 500 ppm NH3, 5% O2
`and 2% water. For the fast SCR reaction studies, catalysts with the
`same loadings as described above were used. But higher space
`velocities were required to achieve smaller conversions since this
`reaction is faster compared to the other reactions. For the Fe-
`zeolite system, shorter catalyst lengths (1 cm of FeZ-11 and 5 mm
`of FeZ-22) were used to achieve lower conversions. Both the
`catalysts had 28 channels. For the Cu-zeolite system, the number
`of channels was reduced from 28 to 9 to achieve high space
`velocity and lower conversion. Catalysts with 2 cm length of CuZ-
`10 and 1 cm length of CuZ-20 were used for this study. The feed
`consisted of 500 ppm NO, 500 ppm NO2, 1000 ppm NH3, 5% O2
`and 2% water. The reaction was studied in the temperatures range
`of 200–550 1C. The starting temperature for this set of experi-
`ments was 200 1C so that there was no solid state ammonium
`nitrate deposition on the catalyst surface. All the gas lines were
`kept heated (T4150 1C) to avoid deposition of ammonium nitrate
`and minimize the adsorption of water and ammonia.
`The NO2 SCR reaction was studied using the same catalysts as
`described earlier for the fast SCR reaction. The feed consisted of
`500 ppm NO2, 500 ppm NH3, 0–5% O2 and 0–2% water. Effects of
`O2 and water were studied in separate experiments and are
`described in following section. Again, the reaction was studied
`in the temperatures range of 200–550 1C to avoid the ammonium
`nitrate deposition ( at lower temperatures) on the catalyst surface.
`
`3. Theoretical background: characteristic times, controlling
`regimes and conversion analyses
`
`3.1. Characteristic times
`
`While the main focus of this work is the investigation of wash-
`coat diffusion limitations in various SCR reactions, the potential
`contribution of other transport processes is also examined. In a
`monolith containing washcoated zeolite catalyst, four types of
`transport processes exist: (1) External mass transfer: This involves
`diffusion of species from gas phase to the surface of washcoat. (2)
`Internal mass transfer (washcoat or pore diffusion): This involves
`
`diffusion of a species in the intercrystalline voids (pores) within the
`washcoat. (3) Intracrystalline diffusion: This involves diffusion
`within the nanopores of the zeolite crystallites. (4) Convective flow:
`This involves the flow of the gas mixture through the channels. The
`catalytic reaction is the fifth process, obviously a chemical process.
`A comparison of characteristic times of each of these individual
`processes gives an insight on the rate limiting process (Bhatia et al.,
`2009). The five characteristic times were compared:
`
` transverse diffusion or external mass transfer time: te ¼ R2
`O1=Df ;
` washcoat diffusion time: td,w ¼ R2
`O2=De;
` intracrystalline diffusion time: td,c ¼ R2
`O3=De,c;
` convection or space time, tc¼L//uS;
` reaction time: tr ¼ ðCi=riÞðRO1=RO2Þ.
`
`Here RO2 is the effective washcoat thickness (or effective
`transverse diffusion length for the washcoat defined as the ratio
`of washcoat cross-sectional area AO2 to the fluid–washcoat inter-
`facial perimeter PO1); De is the effective diffusivity of the reactant
`in the washcoat; L is the channel length, /uS is average gas
`velocity in the monolith channel; RO1 is the effective transverse
`diffusion length (defined as the ratio of channel cross-sectional
`area (AO1) to PO1); RO3 is the zeolite crystallite particle radius; Df
`is the diffusivity of a reacting species in the gas phase and De,c is
`the effective diffusivity in the zeolite crystallite. Diffusion within
`the zeolite crystallite level is included for completion, notwith-
`standing the uncertainty about the value of the diffusion coeffi-
`cient in the zeolite channels and the location of the active sites
`within the pores or on the surface of the crystallites (Metkar et al.,
`2011). We represent a shape of a typical washcoat in a wash-
`coated monolithic channel
`in Fig. 1a and b. There are two
`additional important parameters. The transverse Peclet number
`is the ratio of characteristic times for the transverse (to con-
`vective flow) gas phase diffusion and convection processes
`(P¼te/tc), and is defined as
`/uS
`P ¼ R2
`O1
`LDf
`
`ð8Þ
`
`The magnitude of P determines the upper bound on conversion
`that can be attained in a monolith (in the external mass transfer
`controlled regime). The Weisz–Prater modulus (C) provides an
`estimate of the extent of washcoat diffusion limitations and is
`defined as
`C ¼ R2
`O2Robs
`DeCi
`
`ð9Þ
`
`where Robs is the observed rate of reaction (washcoat volume
`basis). If C41, then washcoat diffusion limitations exist while if
`C51, then diffusion limitations are negligible. Corresponding
`estimates of all these parameters for various SCR reactions are
`shown in Tables 1–3. These findings are discussed in the later
`section of this study where we discuss our experimental findings.
`In most of our experiments, we used small lengths of monolith
`pieces (0.5–2 cm). For this reason, the entrance length effects
`associated with the development of the axial flow velocity profile
`need to be included in the analysis. If the flow is fully developed
`within a short fraction of total length, then the entrance length
`effect can be assumed to be negligible. Ramanathan et al. (2003)
`presented correlations for developing flow in catalytic monoliths.
`Using those correlations, Clayton et al. (2008) found that the flow
`was fully developed within the first 3% of the monolith channel
`used in their study. In the current study, we used the correlation
`for local Sherwood number as a function of axial distance for
`simultaneously developing flow in a circular channel. This corre-
`lation appeared in a recent study (Gundlapally and Balakotaiah,
`
`BASF-2031.003
`
`

`
`P.S. Metkar et al. / Chemical Engineering Science 66 (2011) 5192–5203
`
`5195
`
`Table 2
`Estimated values of various characteristic times and the dimensionless numbers P
`(transverse Peclet number) and C (Weisz–Prater modulus) in the temperature
`range of 200–500 1C for the fast SCR reaction (FeZ-22 catalyst).
`
`Temperature
`(1C)
`
`tc  103
`(s)
`
`te  103
`(s)
`
`td,w  103
`(s)
`
`tr  103
`(s)
`
`P
`
`C
`
`200
`250
`300
`350
`400
`450
`500
`
`5.87
`5.31
`4.84
`4.45
`4.12
`3.84
`3.59
`
`1.73
`1.46
`1.24
`1.07
`0.95
`0.84
`0.75
`
`3.66
`3.08
`2.64
`2.29
`2.01
`1.78
`1.59
`
`25
`7.5
`4.8
`3.6
`2.9
`2.5
`2.2
`
`0.30
`0.28
`0.26
`0.24
`0.23
`0.22
`0.21
`
`1
`2.8
`3.8
`4.4
`4.8
`4.9
`4.8
`
`Table 3
`Estimated values of various characteristic times and the dimensionless numbers P
`(transverse Peclet number) and C (Weisz–Prater modulus) in the temperature
`range of 200–500 1C for the NO2 SCR reaction (FeZ-22 catalyst).
`
`Temperature
`(1C)
`
`tc  103
`(s)
`
`te  103
`(s)
`
`td,w  103
`(s)
`
`tr  103
`(s)
`
`P
`
`C
`
`200
`250
`300
`350
`400
`450
`500
`
`5.87
`5.31
`4.84
`4.45
`4.12
`3.84
`3.59
`
`2.25
`1.89
`1.62
`1.4
`1.2
`1.1
`0.96
`
`4.77
`4.01
`3.42
`2.96
`2.59
`2.29
`2.04
`
`32
`12.3
`7.4
`3.6
`2.7
`2.5
`2.4
`
`0.38
`0.36
`0.33
`0.31
`0.30
`0.28
`0.27
`
`1
`2.2
`3.2
`5.6
`6.6
`6.3
`5.9
`
`P = 0.05
`P = 0.1
`P = 0.2
`P = 0.4
`
`10-3
`
`10-2
`z
`
`10-1
`
`100
`
`100
`
`90
`
`80
`
`70
`
`60
`
`50
`
`40
`
`30
`
`20
`
`10
`
`0
`10-4
`
`Sh (z)
`
`Fig. 2. Computed diagram showing the dependence of the local Sherwood number
`on dimensionless axial position in a circular channel for simultaneously develop-
`ing flow for different values of transverse Peclet number (P¼0.05–0.4).
`
`2011) and is given as follows:
`
`ShðzÞ ¼ 4:364þ 0:98Sc1=6ðP=zÞ
`1þ0:512ðP=zÞ1=2
`
`ð10Þ
`
`where Sh(z) is the local Sherwood number, Sc is the Schmidt number
`and z is the dimensionless coordinate along the length of the
`channel (z¼x/L). We considered different values of transverse Peclet
`numbers (P) in the range of 0.05–0.4 (Tables 1–3) as per our
`experimental conditions. Fig. 2 shows the dependence of the local
`Sherwood number on the dimensionless channel length. From the
`results obtained, it is seen that the flow is almost fully developed (as
`Sh(z) reaches its asymptotic value) within 5% channel length for the
`value of P up to 0.2. For high space velocity experiments (P values of
`
`A♝2
`
`A♝1
`
`P♝1
`
`Washcoat
`
`Cordierite wall
`
`Fig. 1. (a) Scanning electron microscope (SEM) image of the washcoated Cu-
`zeolite monolith catalyst. Washcoat shape was found to be nearly circular with
`more deposition at the corners compared to that in the center. (b) Schematic
`diagram illustrating various length scales in a typical washcoated monolithic
`channel.
`
`Table 1
`Estimated values of various characteristic times and the dimensionless numbers P
`(transverse Peclet number) and C (Weisz–Prater modulus) in the temperature
`range of 200–575 1C for the standard SCR reaction (Fe-Z-XX catalyst).
`
`Temperature
`(1C)
`
`tc  103
`(s)
`
`te  103
`(s)
`
`td,w  103
`(s)
`
`tr  103
`(s)
`
`P
`
`C
`
`200
`250
`275
`300
`325
`350
`425
`500
`575
`
`5.70
`5.15
`4.92
`4.70
`4.50
`4.32
`3.86
`3.49
`3.18
`
`1.73
`1.46
`1.33
`1.24
`1.19
`1.07
`0.89
`0.75
`0.64
`
`5.73
`4.83
`4.42
`4.12
`3.94
`3.55
`2.94
`2.47
`2.12
`
`184.5
`74
`46.2
`37.5
`22.3
`16.7
`5.54
`3.48
`2.86
`
`0.30
`0.28
`0.27
`0.26
`0.25
`0.25
`0.23
`0.21
`0.20
`
`0.17
`0.35
`0.53
`0.61
`0.97
`1.17
`2.92
`3.92
`4.34
`
`BASF-2031.004
`
`

`
`5196
`
`P.S. Metkar et al. / Chemical Engineering Science 66 (2011) 5192–5203
`
` Washcoat diffusion regime: The monolith is said to be in washcoat
`diffusion controlled regime if Rw40.9Rt. In this regime, the
`conversion in the monolith is strongly influenced by changes in
`the washcoat properties.
` External mass transfer regime: The monolith is in external mass
`transfer regime if Rext40.9Rt.
`
`We have used these criteria along with simplified first order
`kinetics (Rate¼kC; k¼1e13nexp(11500/T)s
` 1) to gain insight
`on the controlling regimes for the Fe- and Cu-zeolite catalysts.
`A typical value of the activation energy for the fast and NO2 SCR
`reactions of 96 kJ/mol was used (Chatterjee et al., 2007; Olsson et al.,
`2008). An estimate of the first order rate constant (k0) was
`determined using the experimentally measured rates (washcoat
`volume basis) obtained during the high space velocity experiments
`(and in the absence of transport limitations). A representative case
`for NO2 SCR reaction is shown in Fig. 3a. Here, we show the effect of
`temperature on the transition between regimes for various values of
`
`Rw = 0.25 Rt
`
`Rw = 0.5 Rt
`
`Washcoat
`Diffusion
`Controlled
`
`External Mass
`Transfer
`Controlled
`
`Reaction
`Controlled
`
`10-6
`
`10-7
`De (m2/s)
`
`10-8
`
`10-9
`
`Mass Transfer
`Controlled Regime
`
`Mixed Regime
`
`Rrxn = 0.1 Rt
`
`Kinetic Regime
`
`10-6
`
`10-7
`De (m2/s)
`
`Rrxn = 0.9 Rt
`
`10-8
`
`10-9
`
`1000
`
`900
`
`800
`
`700
`
`600
`
`500
`
`400
`
`300
`
`200
`
`100
`
`0
`10-5
`
`350
`
`300
`
`250
`
`200
`
`150
`
`100
`
`50
`10-5
`
`Temperature (°C)
`
`Temperature (°C)
`
`Fig. 3. (a) Diagram showing the effect of temperature and effective washcoat
`diffusivity on regime transition in a catalytic monolith. (b) Below the lower curve
`(Rrxn¼0.9Rt), reaction controlled (kinetic) regime is dominant. Above the upper
`curve (Rrxn¼0.10 Rt), mass transfer (external and internal) controlled regime is
`dominant.
`
`0.3–0.4), the length required for the complete flow development
`increased but did not exceed 8% of the total channel length. As the
`flow was found to be fully developed within a very short channel
`length, we neglected the entrance length effects (assuming fully
`developed flow) in our further calculations.
`
`3.2. Controlling regimes in catalytic monoliths
`
`Mass transfer limitations are of prime importance in all the
`catalytic reactions. While the experimental study on catalysts with
`different washcoat thickness is helpful in investigating washcoat
`diffusion limitations, external mass transfer limitations should also
`be considered. A recent study (Joshi et al., 2010) showed the effect of
`temperature on transition between various controlling regimes in
`catalytic monoliths (from kinetic to washcoat diffusion to external
`mass transfer) for the case of an isothermal monolith. They used a
`resistances-in-series approach to determine the controlling regime.
`The following resistances were considered in their study:
`
` Fluid phase film or external mass transfer resistance (Rext):
`
`Rext ¼ 1kmeðzÞ ¼ 4RO1
`ð11Þ
`SheðzÞDf
`where She is the external Sherwood number and kme is the
`external mass transfer coefficient (from the bulk of the fluid–
`washcoat interface). If the entrance effects are negligible, a
`constant value of She(z) may be used in Eq. (11), e.g. She(z)¼4.36
`for a circular channel.
` Internal mass transfer (washcoat diffusional) resistance (Rw):
`Rw ¼ 1
`RO2
`¼
`ð12Þ
`ShiðfÞDe
`kmi
`where kmi is the internal mass transfer coefficient between the
`interior of the washcoat and the fluid–washcoat interface, Shi
`is the internal Sherwood number and for linear kinetics, it may
`be expressed as
`
`ShiðfÞ ¼ Shi1 þ Lf2
`1þLf2
`(As we show below, we approximate the kinetics of the exam-
`ined reactions as first order.) Values of ShiN and L depend upon
`the washcoat geometries (Joshi et al., 2010). Here, the inner
`boundary of the washcoat was assumed to be circular in shape
`and the washcoat has a constant thickness. This gives values for
`ShiN¼3 and L¼0.32. The washcoat Thiele modulus (f) is
`
`ffiffiffiffiffiffiffiffiffis
`defined as
`kðTÞ
`De
`
`ð13Þ
`
`ð14Þ
`
`f ¼ RO2
`
`where k is the first order rate constant (having units of inverse
`time) based on washcoat volume.
`
` Reaction resistance (Rrxn):
`Rrxn ¼
`
`1
`kðTÞRO2
`
`ð15Þ
`
`The sum of all these resistances gives the total resistance (Rt)
`for the conversion of reactants to products:
`1
`Rt ¼ 4RO1
`þ RO2
`þ
`SheðzÞDf
`ShiðfÞDe
`kðTÞRO2
`In order to determine the controlling regime, Joshi et al.
`designed some practical criteria described as follows:
` Kinetic regime: The monolith is said to be in the kinetic regime
`(or conversion in the monolith is mainly determined by the
`kinetics) if Rrxn40.9Rt.
`
`ð16Þ
`
`BASF-2031.005
`
`

`
`P.S. Metkar et al. / Chemical Engineering Science 66 (2011) 5192–5203
`
`5197
`
`the reactant effective diffusivity in the washcoat. From Fig. 3a, it is
`clear that very low values of effective washcoat diffusivities (De) are
`required to attain purely washcoat diffusion controlled regime.
`However, washcoat diffusion is seen to be practically important
`for most of the temperatures as the reaction system is found to be in
`mixed regime. Also, we show the curves that demarcate the regions
`in which the washcoat diffusion resistance is an important fraction
`of the total resistance (Rw40.25Rt and Rw40.5Rt). In Fig. 3b, we
`show two curves for the cases when reaction resistance is a
`resistance (Rrxn¼0.10Rt, and
`significant
`fraction of
`the total
`Rrxn¼0.90Rt). For the case of Rrxn40.90Rt (below the lower curve),
`the reactant conversion is controlled by the kinetics. If we want to
`study intrinsic kinetics, reaction should be carried out in this regime.
`For Rrxno0.10Rt (above the upper curve), the reaction resistance is
`negligible and the reactant conversion is controlled by internal and
`external mass transfer. If experiments are carried out in this regime,
`then the data should not be used to extract kinetics. To attain purely
`external mass transfer controlled regime, very high temperatures
`are required.
`
`3.3. Effect of washcoat properties on conversions in an isothermal
`monolith
`
`The value of total resistance (Rt) obtained in Eq. (16) can be
`used to determine the apparent mass transfer coefficient (km,app)
`in washcoated monoliths:
`km,app ¼ 1
`Rt
`
`ð17Þ
`
`Df /De = 10
`Df /De = 100
`Df /De = 1000
`
`0
`
`100
`
`200
`
`300
`
`600
`500
`400
`Temperature (°C)
`
`700
`
`800
`
`900 1000
`
`1
`0.9
`0.8
`0.7
`0.6
`0.5
`0.4
`0.3
`0.2
`0.1
`0
`
`Conversion
`
`Fig. 4. Diagram showing the effect of temperature on conversion and regime
`transitions in a catalytic monolith for three values of effective diffusivity (hollow
`circle: Rrxn¼0.9Rt, dark circle: Rrxn¼0.1Rt, rectangle: Rext¼0.9Rt).
`
`Fig. 4 where we indicate the temperatures at which transitions
`occur between the kinetic regimes (Rrxn40.90Rt), mixed (mass
`transfer) regime (Rrxno0.10Rt) and the purely external mass
`transfer controlled regime (Rext40.9Rt). Three different values
`of the Df/De ratios (10, 100 and 1000) are used. From the results, it
`is clear that the reaction was in kinetic regime for temperatures
`below 190 1C for all the three cases. The increasing Df/De ratio
`shifts the reaction regime temperature towards the left even
`though the shift is not large. For Df/De¼10, the effect of washcoat
`diffusion limitations is found to be negligible as the difference in
`transition temperatures from mixed mass transfer regime to
`external mass transfer regime was negligible. This temperature
`window increases further for the case of Df/De¼100 and thus
`washcoat diffusion limitations become important for wide tem-
`perature range. Also, the temperature required to attain pure
`external mass transfer regime shifts towards the right. For the case
`of Df/De¼1000, the reaction is found to be in the mixed mass
`transfer regime for most of the temperatures and washcoat diffusion
`limitations become very important. The temperature required to
`achieve pure external mass transfer regime increases further.
`The results shown in Fig. 4 are for a representative case of the
`NO2 SCR reaction which shows the effect of effective washcoat
`diffusivities on the regime transition for one fixed value of catalyst
`loading. If the catalyst loading is varied, then the temperatures
`required to achieve the corresponding regimes will shift accordingly.
`From this figure, it is seen that the reaction is in kinetic regime for a
`very narrow temperature range. If the catalyst loading is decreased
`(metal content or washcoat loading), then the temperature range for
`the kinetic regime could be broadened. The above case for NO2 SCR
`should be considered as a representative case of all the SCR
`reactions. Clearly, a more detailed kinetic model would be required
`for the accurate prediction of all the regimes. These theoretical
`calculations are revisited in the next section of this study where we
`present our experimental observations.
`The experimental study on catalysts with the same washcoat
`loadings but different washcoat thicknesses is useful to investigate
`washcoat diffusion limitations. Here we use a simple isothermal
`monolith model based on linear kinetics to study the above
`representative case. The conversion of the limiting reactant (say
`NO for standard SCR case) is calculated using Eq. (19). The results
`are summarized in Fig. 5, which shows the plots obtained for
`conversion vs temperature (w vs T) for the three cases of assumed
`values of effective washcoat diffusivities (Df/De¼10, 100 and 1000).
`The solid lines represent catalysts with the thinner washcoat while
`
`
`
`!
`
`
`w ¼ 1a1 exp
`
`For an isothermal monolith, value of km,app can be used to
`
`calculate the conversion of a limiting reactant using following
`equation (Joshi et al., 2010):
`km,appL
`w ¼ 1a1 exp
`RO1/uS
`Substituting for km,app and writing it in terms of dim