On kinetic modeling of change in active sites upon hydrothermal aging of Cu-SSZ-13
Rohil Daya, Saurabh Joshi, Jinyong Luo, Rama Krishna Dadi, Neal Currier, Aleksey Yezerets
Public
09-18-2019
CLEERS 2019 Workshop
2Public
Outlineβͺ Motivation and Background
βͺ Modification of reaction rate forms to account for active site dynamics
βͺ Developing constitutive relations for Hydrothermal Aging of Cu-SSZ-13
βͺ Summary and Future Work
3Public
Motivation β Tailpipe NOx and Durability Requirements
Future Application Challenges
2024+
Challenge
Real World 90%NOx
Add heat/CDA/ bypass
Sustained LT
Cold start
GHG Reduction
Engine efficiency
Electrification
Fuel type
β¦
Emission Warranty Extension
Forecast 2022 2024
HD 5Y 350K miles 5Y 435K miles ( maybe longer)
MD 5Y 150K miles 5Y 185K miles ( maybe longer)
Overall approach to meet these challenges
includes
a) Adaptable Controls
b) Improved understanding and modeling
of real world aging
c) Improved Component durability (higher
resistance to real world aging)Class 2022 2027 (forecast)
HD 5Y 350K miles 14Y 800K miles
MD 5Y 150K miles 14Y 400K miles
4
Two chemically distinct monomeric Cu2+ active sites in Cu-SSZ-13; typically solvated and mobile
under reaction conditions; undergo redox during SCR (Cu1+ observed during standard SCR)
Mild-hydrothermal aging leads to
decrease in number density of ZCuOH sites (πΊππΆπ’ππ»),and concomitant increase in number density of Z2Cu
sites (πΊπ2πΆπ’) [1]. Severe HTA leads to dealumination,
increase in CuOx, and eventually structural collapse
Public
Active Sites in Cu-SSZ-13 and influence of aging
Hydrothermal Aging Sulfur Poisoning
ZCuOH
Z2Cu
[1] Luo, J., An, H., Kamasamudram, K., Currier, N., Yezerets, A., Watkins, T., & Allard, L.
(2015). Impact of Accelerated Hydrothermal Aging on Structure and Performance of Cu-SSZ-
13 SCR Catalysts. SAE International Journal of Engines, 8(3), 1181-1186.
Sulfur selectively poisons Copper sites.
ZCuOH stores more sulfur and is harder to
regenerate relative to Z2Cu [2]
[2] Shih, A. J., Khurana, I., Li, H., GonzΓ‘lez, J., Kumar, A., Paolucci, C., ... & Yezerets, A.
(2019). Spectroscopic and kinetic responses of Cu-SSZ-13 to SO2 exposure and implications
for NOx selective catalytic reduction. Applied Catalysis A: General, 574, 122-131.
5
Monolith reactors can be modeled as plug flow reactors. Equations below
are sufficient to describe SCR behavior under realistic operating conditions
Public
Classical Conservation Equations for Monolith
Reactors
ππππππ
= βπππ,ππ πππ
(ππ β ππππ)ππ,πππ
πͺπ= βπ«πππ
πππππππππ
π΄πππ½πππ
= ππ,πππ
Convection Rate = Diffusion Rate Surface Coverage ConservationReaction Rate = Diffusion Rate
ππ,π ππ β ππππ = βπ«ππππππππππ
ππ π = πsubject to
[3] Copeland, C., Pesiridis, A., Martinez-Botas, R., Rajoo, S., Romagnoli, A., & Mamat, A.
(2014). Automotive Exhaust Waste Heat Recovery Technologies. In Automotive Exhaust
Emissions and Energy Recovery. Nova Science Publishers.
Copeland et al. [3]
0 200 400 600 800 1000 1200
Cu
mu
lati
ve N
Ox
[g]
Time [s]
Hot FTP Cumulative NOx
SCR-In NOxSCR-Out ExpSCR-Out Model
0 200 400 600 800 1000 1200
NH
3[p
pm
]
Time [s]
Hot FTP NH3 Slip
SCR-Out Exp
SCR-Out Model
6
βͺ Aging induces changes in catalyst state. These changes can be modeled in a classical framework
by deriving:
a) Constitutive relations/rate expressions for quantitative changes of all chemically distinct active sites
b) Intrinsic, site specific turnover rates for conversion of gas species on each active site
βͺ In this representation, the state and behavior of each active site is defined completely by a set of
variables and constants with respect to aging
βͺ This entails a combination of reaction engineering and more fundamental approaches incorporating
DFT predictions and microkinetics
βͺ In the next few slides, we will show a first effort at developing constitutive relations for the change in
active sites induced by hydrothermal aging of Cu-SSZ-13 Public
Incorporating Aging in Conservation Equations
π΄πππ½πππ
= ππ,πππ π΄πβ² (ππππ, π»πππ, πΊπΆπβ¦ )
ππ½πππ
= ππ,πππβ²(ππππ, π»πππ, πΊπΆπβ¦ )
Fixed Age Variable Age
Z2Cu
Variable : πΊπ2πΆπ’ (ππ’ππππ ππππ ππ‘π¦)
Constants (if fixed turnover rate) : π΄ππ»3π΄ππ π2πΆπ’ ,
βπππ»3π΄ππ π2πΆπ’ , βπ»ππ»3π΄ππ π2πΆπ’ , π΄ππΆπ π2πΆπ’ , πΈπππΆπ π2πΆπ’β¦
ZCuOH
Variable : πΊπ2πΆπ’ (ππ’ππππ ππππ ππ‘π¦)
Constants (if fixed turnover rate) : π΄ππ»3π΄ππ ππΆπ’ππ» ,
βπππ»3π΄ππ ππΆπ’ππ» , βπ»ππ»3π΄ππ ππΆπ’ππ» , π΄ππΆπ ππΆπ’ππ» , πΈπππΆπ ππΆπ’ππ» β¦
7Public
Hydrothermal Aging Model Approach
NH3 storage kinetics on individual sites
β’ Site isolation approach
β’ 550Β°C-4h H-SSZ-13 data used for BrΓΈnsted acid site kinetics
β’ 750Β°C-4h Cu-SSZ-13 data used for lumped Copper site kinetics.
β’ 600Β°C-2h low temperature (
8Public
Active Sites for NH3 Adsorption on H-SSZ-13βͺ It has been shown that on H-SSZ-13, NH4
+ cations are NH3 βsolvated below 400Β°C, based on in-
situ vibrational spectroscopy and DFT simulations [4-5]
[4] Giordanino, F., Borfecchia, E., Lomachenko, K. A., Lazzarini, A., Agostini, G., Gallo, E., ... & Lamberti, C.
(2014). Interaction of NH3 with Cu-SSZ-13 catalyst: a complementary FTIR, XANES, and XES study. The
journal of physical chemistry letters, 5(9), 1552-1559.
[5] Li, S., Zheng, Y., Gao, F., Szanyi, J., & Schneider, W. F. (2017). Experimental and computational
interrogation of fast SCR mechanism and active sites on H-Form SSZ-13. ACS Catalysis, 7(8), 5087-5096.
DFT Optimized Structures [5]
Strong-H Weak-H
βͺ Furthermore, in-house experimental data for H-SSZ-13 shows indirect evidence of second order adsorption,
consistent with solvation of NH4+ ions by NH3 . Thus, NH3 adsorption on BrΓΈnsted acid sites modeled as a
Type-II BET isotherm
ππβππ» = ππβππ»ππ.ππ»4πΊπ»
ππ»3 + ππ» βΏ π.ππ»4ππβππ» = ππβππ»π¦ππ»3πππ»πΊπ»
ππ»3 + π.ππ»4 βΏ π.ππ»4. ππ»3ππβππ» = ππβππ»π¦ππ»3ππ.ππ»4πΊπ»ππβππ» = ππβππ»ππ.ππ»
4.ππ»
3πΊπ»
9Public
Second Order NH3-Adsorption on H-SSZ-13βͺ NH3 release during isothermal adsorption and TPD can be described quantitatively using the Type-II BET
adsorption model
βͺ Results from typical first order Langmuir/Temkin adsorption model show that while such a model can capture
the desorption peak with similar accuracy, NH3 release isothermal NH3 adsorption cannot be described
accurately
550Β°C-4h H-SSZ-13 NH3 TPD 550Β°C-4h H-SSZ-13 Total Storage
10Public
Additional Active Sites for NH3 Adsorption on Cu-SSZ-13In addition to the Strong-H and Weak-H sites, two additional sites considered for NH3 adsorption:
Copper Sites (Temkin isotherm model)
Physisorbed NH3 (to account for increased storage at high feeds and low temperatures. Langmuir
isotherm model)
ππβπΆπ’ = ππβπΆπ’ππΆπ’.ππ»3πΊπΆπ’.ππ»3
ππ»3 + πΆπ’ βΏ πΆπ’.ππ»3ππβπΆπ’ = ππβπΆπ’π¦ππ»3ππΆπ’πΊπΆπ’.ππ»3
ππβπ = ππβπππ.ππ»3πΊπ
ππ»3 + π βΏ π.ππ»3ππβπ = ππβππ¦ππ»3πππΊπ
ZCu, Z2Cu and ZCuOH sites lumped
into global Copper site, due to
inability of NH3-TPD to resolve
individual copper sites
Isolated Cu2+ positions from AIMD [6]
[6] Paolucci, C., Di Iorio, J. R., Ribeiro, F. H., Gounder, R., & Schneider, W. F. (2016). Catalysis science of NOx selective catalytic
reduction with ammonia over Cu-SSZ-13 and Cu-SAPO-34. In Advances in Catalysis (Vol. 59, pp. 1-107). Academic Press.
11Public
Copper and Physisorbed Site Kinetics on Cu-SSZ-13
βͺ NH3-TPD at 750Β°C-4h shows a primary release feature at 290Β°C, with a minor shoulder at 400Β°C
βͺ Significant low temperature isothermal desorption attributed to a global physisorbed site. Isothermally
desorbed NH3 shows non-monotonic trend with increased temperature
Age : 600Β°C-2h
Experiment : 1h isothermal
adsorption followed by 40
minutes isothermal desorption
12Public
Site-Specific Adsorption-Desorption Dynamicsβͺ Quantitative predictions of the influence of catalyst temperature on storage of individual sites
βͺ Steady-state total storage on 750Β°C-4h aged Cu-SSZ-13 compares well with model predictions
Age : 600Β°C-2h
Experiment : 1h isothermal
adsorption followed by 40
minutes isothermal desorption
13Public
NH3 Adsorption Equilibrium Thermodynamics
[7] Paolucci, C., Parekh, A. A., Khurana, I., Di Iorio, J. R., Li, H., Albarracin Caballero, J. D., ... & Ribeiro, F. H.
(2016). Catalysis in a cage: condition-dependent speciation and dynamics of exchanged Cu cations in SSZ-13
zeolites. Journal of the American Chemical Society, 138(18), 6028-6048.
βͺ Estimated adsorption enthalpies and entropic penalties are aligned with DFT/AIMD predicted values in
literature
βͺ Entropic penalties in line with the notion of mobile NH3 complexes on zeolite surface
[5] Li, S., Zheng, Y., Gao, F., Szanyi, J., & Schneider, W. F. (2017). Experimental and computational interrogation of fast SCR
mechanism and active sites on H-Form SSZ-13. ACS Catalysis, 7(8), 5087-5096.
Model ParametersRecent DFT/AIMD Predictions [5, 7-8]
[8] Li, H., Paolucci, C., & Schneider, W. F. (2018). Zeolite adsorption free energies from ab initio
potentials of mean force. Journal of chemical theory and computation, 14(2), 929-938.
14Public
Modeling the change in NH3-TPD with hydrothermal
agingβͺ NH3-TPD experiments indicate a monotonic decrease in high temperature peak and increase in low
temperature peak with increased aging [9]
βͺ This behavior was modeled with:
a) Fixed turnover rates (identified previously)
b) Increased NH3 storage ability on Copper sites with aging
c) Decreased number density of BrΓΈnsted acid sites with aging
[9] Luo, J., Gao, F., Kamasamudram, K., Currier, N., Peden, C. H., & Yezerets, A.
(2017). New insights into Cu/SSZ-13 SCR catalyst acidity. Part I: Nature of acidic
sites probed by NH 3 titration. Journal of Catalysis, 348, 291-299.
15Public
Tracking Storage Capacity on Copper Sites and
Number Density of BrΓΈnsted Acid Sitesβͺ Increase in NH3 storage on copper sites is linearly related to the loss of BrΓΈnsted acid sites
βͺ The loss of BrΓΈnsted acid sites with aging is used to develop a constitutive relationship for the
evolution of these sites, yielding the hydrothermal aging equation
NH3 Storage Ability on
Copper Sites
Number Density of
BrΓΈnsted acid Sites
16Public
Hydrothermal Aging Equation
π
π΄π―ππππ= π. ππ β ππππππ β
ππππππ
πΉπ»πππππππ + π
Ratio-based calculator reported previously [10] Identical deactivation energies from two
different derivation methods
βͺ A linear relationship is found between the inverse BrΓΈnsted acid site density and aging time,
implying a second order Arrhenius rate. The rate constant and activation energy are derived
Arrhenius Plot
[10] Luo, J., Kamasamudram, K., Currier, N., & Yezerets, A. (2018). NH3-TPD methodology for quantifying
hydrothermal aging of Cu/SSZ-13 SCR catalysts. Chemical Engineering Science, 190, 60-67.
17Public
Model Limitations β Copper Site Deconvolution
βͺ NH3 adsorption at 200Β°C leads to replacement of approximately two framework oxygen atoms on
Z2Cu sites and approximately one framework oxygen atom on ZCuOH sites [9]
π΄πͺπ.π΅π―π = ππ΄πππͺπβ² + ππ΄ππͺππΆπ―
β²
π΄ππͺππΆπ―β²
π΄πππͺπβ² + π΄ππͺππΆπ―
β² = π. ππ ππ πππΒ°πͺ β πππ π΄πππͺπ
β²
π ππππ= β
π π΄ππͺππΆπ―β²
π ππππ
βπ π΄π―
π ππππ= β
π π΄ππͺππΆπ―β²
π ππππ+ π. ππ
π π΄πππͺπβ²
π ππππ
Known from NH3-TPD model
βͺ Model does not explicitly predict the evolution of individual Copper sites with hydrothermal aging.
However, combining model results with other characterization and experimental data can provide
some insight
Reducible Cu2+ densities
βͺ Additional characterization data on same catalyst :
H2-TPR on degreened catalyst [9]
βͺ Correlating BrΓΈnsted acid site loss rate with
individual copper site change rate :
NO + NH3 titration
experiments
[9] Luo, J., Gao, F., Kamasamudram, K., Currier, N., Peden, C. H., & Yezerets, A.
(2017). New insights into Cu/SSZ-13 SCR catalyst acidity. Part I: Nature of acidic
sites probed by NH 3 titration. Journal of Catalysis, 348, 291-299.
18Public
Summary and Future Workβͺ Modeling of SCR catalyst aging is necessary to meet the next phase of tailpipe regulations and
durability requirements
βͺ If successfully developed and integrated with performance predictions, these models have
significant applications in improved understanding and design of SCR catalysts, along with better
catalyst health monitoring and control
βͺ These models can be enabled by intrinsic micro-kinetics for species conversion on each active site
and experimental quantification of active site evolution with aging. For instance, the loss of
BrΓΈnsted acid sites with hydrothermal aging can be expressed by :
βͺ Constitutive relations must be discovered for the response of all chemically distinct active sites to
different SCR degradation mechanisms
βͺ Once the catalyst state is known, the performance can be estimated using site-specific turnover
rates identified through experimental and theoretical understanding
π
π΄π―ππππ= π. ππ β ππππππ β
ππππππ
πΉπ»πππππππ + π
1919
20
Backup Slides
Public
21Public
Transient and Steady-State Coverages on H-SSZ-13βͺ The surface coverages are plotted as a function of time (transient TPD) and temperature (steady-state).
Steady-state coverages compare qualitatively well with computational results in [5]
[5] Li, S., Zheng, Y., Gao, F., Szanyi, J., & Schneider, W. F. (2017). Experimental and computational interrogation of fast SCR
mechanism and active sites on H-Form SSZ-13. ACS Catalysis, 7(8), 5087-5096.
Surface Coverages in [5]
550Β°C-4h H-SSZ-13 NH3 TPD550Β°C-4h H-SSZ-13 Total Storage
22Public
Experimental Data : Jinyong Luo
Transient and Steady-State Coverages on Cu-SSZ-13βͺ The surface coverages are plotted as a function of time (transient TPD) and temperature (steady-state)
23Public
Validation of modeling approach with steady-state
Storage measurementsβͺ Evolution of total NH3 storage with hydrothermal aging can be understood using the model
βͺ With hydrothermal aging, the significance of BrΓΈnsted acid sites as a reservoirs of NH3 decreases
Surprising experimental result at 450Β°C being
investigated further
24Public
Analytical Expressions for Total and Site Specific Storageβͺ The equilibrium constants on each site can be utilized to calculate site specific surface coverage using
Langmuirβs adsorption isotherm
πΏπππππ’ππ/ππππππ ππ’πππππ πΆππ£πππππ ββ π½π¬ππ(ππ΅π―π , π») =π²π¬ππππ΅π―π
π+π²π¬ππππ΅π―ππ = πΆπ’ π ππ‘ππ , πβπ¦π π ππ‘ππ
βͺ Once the surface coverage on each of the sites is known, the total ammonia storage can be calculated as
follows:
ππ΅π―π = π΄πΎπ΅π―ππΎπ³
πππ΄πͺπ.π΅π―ππ½π¬ππͺπ + π΄π― π½π¬πππ΅π―π
+ ππ½π¬πππ΅π―ππ΅π―π+ π΄ππ½π¬ππ·
Where ππ΅π―π is the NH3 storage in g/Lcat, π΄πΎπ΅π―π is the molecular weight of NH3 in kg/mol, πΎπ³ is washcoat loading in
g/Lcat, ππ is washcoat density in kg/m3 (666 for DW3136 family), π΄π mol/m
3washcoat and π½π¬ππ is the equilibrium surface
coverage of active site π
βͺ In order to estimate NH3 storage as a function of aging time and temperature, π΄π as a function of aging time and temperature must be calculated, and this is the objective of the aging model
π΅πΈπ ππ’πππππ πΆππ£πππππ ββ π½π¬πππ΅π―π(ππ΅π―π , π») =π²π¬ππΊπ―ππ΅π―π
π+π²π¬ππΊπ―ππ΅π―π+π²π¬ππΊπ―π²π¬ππΎπ―ππ΅π―ππ ; π½π¬πππ΅π―ππ΅π―π(ππ΅π―π , π») =
π²π¬ππΊπ―π²π¬ππΎπ―ππ΅π―ππ
π+π²π¬ππΊπ―ππ΅π―π+π²π¬ππΊπ―π²π¬ππΎπ―ππ΅π―ππ
25Public
Model Limitations β Severe Hydrothermal Aging
βͺ Beyond a threshold aging, there is a change in
NH3 storage energetics on Copper sites,
characterized by a shift in the low temperature
release peak from 290Β°C to 265Β°C
βͺ Systematic analysis of NH3 TPD profile shows
that are at least two different phases of
hydrothermal aging:
Phase I β Drop in ratio of high to low temperature
peaks, but negligible loss of Cu2+ (transformation of
ZCuOH to Z2Cu)
Phase II β Drop in Cu2+ sites. BrΓΈnsted acid sites
negligible. Change in energetics of low temperature
sites. Loss of NH3 storage
26Public
Cu2+ site Distribution as a function of Hydrothermal
Age
Key Model Prediction : In the
degreened state, 30-35% of the
Copper does not store NH3, and
must therefore not exist as
isolated Cu2+
π2πΆπ’ π€π‘.% =πΊπ
2πΆπ’ β πππΆπ’ β πππ·πΊ β 610.237
ππΆπ’ππ» π€π‘.% =πΊππΆπ’ππ» β πππΆπ’ β ππ
π·πΊ β 610.237
Where πππΆπ’ is the molecular weight of Cu in
g/mol, ππ is ratio of washcoatvolume to reactor volume
and π·πΊ is dry gain in g/in3