See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/345811530

A modeling and experimental study on hydrothermal aging deactivation of NO

oxidation activity on Pt-Pd catalyst

Article in Applied Catalysis B: Environmental · October 2020

DOI: 10.1016/j.apcatb.2020.119655

CITATIONS READS

0 29

9 authors, including:

Rama Krishna Dadi Rohil Daya

Cummins Inc. Cummins Inc.
11 PUBLICATIONS 38 CITATIONS 9 PUBLICATIONS 15 CITATIONS

SEE PROFILE SEE PROFILE

Ashok Kumar Saurabh Y. Joshi

Open University Malaysia Cummins Inc.
58 PUBLICATIONS 720 CITATIONS 25 PUBLICATIONS 373 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Micro-kinetic modeling. View project

TWC Modeling View project

All content following this page was uploaded by Rama Krishna Dadi on 13 November 2020.

The user has requested enhancement of the downloaded file.

 Applied Catalysis B: Environmental 283 (2021) 119655

Contents lists available at ScienceDirect

Applied Catalysis B: Environmental

journal homepage: www.elsevier.com/locate/apcatb

A modeling and experimental study on hydrothermal aging deactivation of

NO oxidation activity on Pt-Pd catalyst
Rama Krishna Dadi *, Rohil Daya , Ashok Kumar , Saurabh Y. Joshi , Hongmei An ,
Michael J. Cunningham , Neal W. Currier 1, Aleksey Yezerets
Cummins Inc., 1900 McKinley Ave, Columbus, IN 47201, USA

A R T I C L E I N F O A B S T R A C T

Keywords: A state-of-the-art Pt-Pd catalyst was hydrothermally treated at different conditions to investigate hydrothermal
Precious metals stability of NO oxidation. NO oxidation reactor data was collected on catalysts that have been hydrothermally
Aging kinetics treated at temperatures ranging from 550 ◦ C until 1100 ◦ C. NO oxidation activity was observed to decrease with
Oxidation catalyst
increase in aging duration at all the aging temperatures. A global kinetic model of NO oxidation was used to
estimate the change in NO oxidation rate when subjected to different hydro-thermal aging conditions. Model
estimated normalized rates at multiple hydro-thermal aging conditions were used to develop a hydro-thermal
aging model of NO oxidation. A deactivation Model with Residual Activity (DMRA) was developed to capture
the impact of hydrothermal aging temperature and duration on NO oxidation activity. A global kinetic model
that accounts for change in PGM oxidation state during NO oxidation has also been presented.

components like Diesel Particulate Filter (DPF) and Selective Catalytic

Reduction (SCR) in Diesel Engine After-Treatment system
1. Introduction
The impact of precious metal dispersion and particle size on NO
oxidation activity has been reported in the past [6–8]. NO oxidation rate
Precious metals like Pt and Pd supported on alumina are widely used
was observed to follow a non-monotonic trend with particle size of Pt-Pd
as oxidation catalysts in industrial applications. These oxidation cata­
based catalysts supported on γ − Al2 O3 [6]. A local maximum in NO
lysts are used to oxidize several chemical species like CO, hydrocarbons,
oxidation rate on Pt based catalysts is typically observed in the average
NO, NH3. Oxidation catalysts are widely utilized in automobile industry
particle size range of 5− 10 nm. Further increase in particle size,
to chemically convert harmful emissions like CO, HCs, NH3 into more
potentially by hydrothermal aging, results in a drop in NO oxidation
benign chemical species. Pt-Pd bimetallic catalysts are one of the good
rate. A number of articles in the past investigated the extent of decline in
candidates for oxidation catalysts because of their superior oxidation
NO oxidation rate with increase in particle size beyond the optimal
performance. However, Pt-Pd catalysts are susceptible to sintering when
value [7,8]. One such study by Liang et al. [8] illustrated the impact of
exposed to high temperatures. Pt is observed to more readily sinter
particle size on NO oxidation rate when a Diesel Oxidation Catalyst was
compared to Pd on Alumina when subjected to hydrothermal aging.
subjected to hydrothermal treatment for different times at an aging
Bimetallic catalysts of Pt and Pd are observed to enhance the thermal
temperature of 700 ◦ C. It was observed in this study that average particle
stability of metallic Pt [1,2]. Research studies in the past investigated
size changes between 10− 17 nm when hydrothermally treated for aging
the oxidation performance of bimetallic catalysts as function of Pt: Pd
durations in the range of 1− 100 h at 700 ◦ C. A correlation between Turn
weight ratio [3–5]. While some HCs exhibit a non-monotonic oxidation
Over Frequency (TOF) and hydrothermal aging duration was reported in
activity with Pt content, NO oxidation activity is observed to mono­
the study. They also pointed out that nitrate chemistry on the surface
tonically increase with increase in Pt content [5]. One of the objectives
affects the rate of deactivation in NO oxidation activity. The afore­
of this work is to investigate hydrothermal stability of NO oxidation for a
mentioned articles focused on the stability of precious metals with hy­
state-of-the-art Diesel Oxidation catalyst (DOC), a Pt-Pd based catalyst.
drothermal exposure. In addition to the stability of precious metals, a
DOCs are typically used to oxidize NO and HCs in diesel engine exhaust.
stable support structure is also paramount for oxidation functionalities.
Oxidation of NO to NO2 results in the effective operation of downstream

* Corresponding author.
E-mail address: [email protected] (R.K. Dadi).
1
Retired from the role at Cummins Inc.

https://doi.org/10.1016/j.apcatb.2020.119655
Received 31 July 2020; Received in revised form 12 October 2020; Accepted 17 October 2020
Available online 22 October 2020
0926-3373/© 2020 Elsevier B.V. All rights reserved.
R.K. Dadi et al. Applied Catalysis B: Environmental 283 (2021) 119655

Nomenclature Yj Molar fraction of chemical species j

Roman Letters Greek Letters

Ψ Normalized NO oxidation activity (-)
Ai Pre-exponential factor of reaction i (mmol
3 .s)
θ Fraction of catalytic sites
DMRA Deactivation Models with Residual Activity
DOC Diesel Oxidation Catalyst Subscripts
DPF Diesel Particulate filter age Hydrothermal Age
Ei Pre-exponential factor of reaction i (kJ/mol) d Deactivation
GPLE General Power Law Expression eq Thermodynamic equilibrium
HC Hydrocarbon i Reaction number
HTA Hydrothermal Aging (e.g. 550 ◦ C–4 h → HTA Temperature- in Inhibition
HTA duration) j Chemical species
ki Reaction rate constant of reaction i (kJ/mol) m PGM in metallic or reduced form
Rg Universal Gas Constant (J/ (mol. K)) Pt Precious metals (Pt or Pd)
SCR Selective Catalytic Reduction Ref Reference Hydrothermal Aging condition (550◦ C–4h)
SPLE Simple Power Law Expression ∞ Asymptotic value
T Temperature (K)
t Time (s)

Sintering of support structure caused due to extreme hydrothermal 2. Experimental

treatment can also lead to loss in NO oxidation activity [7]. Catalysts
studied in this work have not been exposed to conditions that may result NO oxidation performance data sets were collected on plug flow
in the collapsing of support structure. reactor using the experimental procedure elaborated previously [18].
( ) Most of the reactor testing data shown in this work was done with a
dΨ E∞
GPLE : = − kd (Ψ − Ψ∞ )n ; Ψ∞ = A∞ exp − (1) GHSV of 80,000 h− 1. A state-of-the-art Diesel Oxidation Catalyst sup­
dt Rg Tage
ported on a cordierite substrate was used for reactor study. Monolithic
catalyst cores (1-inch diameter by 3-inch length) were used for reactor
dΨ
SPLE : = − kd (Ψ)n (2) testing. The catalyst cores were hydrothermally aged in a base feed
dt
mixture containing 10 % O2; 7 % H2O and 8 % CO2 in N2 prior to the
Numerous studies in the literature presented sintering models for Pt- experiments. Subsequently, the catalysts cores were cooled down to
Pd catalysts [9–12]. Bartholomew et al. [11] elucidated the necessity of 150 ◦ C in the base feed and NO was introduced in the feed. Steady-state
using Deactivation Models of Residual Activity (DMRA) or General NO oxidation flow reactor tests were performed at different catalyst
Power Law Expressions (GPLE) to explain several supported metal temperatures, starting from 150 ◦ C till 450 ◦ C with an increment of
dispersion data sets that stabilize at non-zero values after long durations 50 ◦ C. Catalyst temperature was held at each temperature for 35 min to
of thermal treatment. Some of the previous studies [11,12] also pointed evaluate NO oxidation activity as a function of temperature.
out that sintering models represented by Simple Power Law Expression
(SPLE), effectively without residual activities, can lead to erroneous 3. Results and discussion
estimation of kinetic parameters of HTA model. DMRA or GPLE models,
shown in Eqn. 1, typically have an additional residual term Ψ∞ as 3.1. Reactor and kinetic models
compared to SPLE models. kd is Arrhenius rate constant for deactivation
rate, n represents the order of thermal deactivation. Researchers A commercial software, AVL BOOST 2019, was used to perform
demonstrated that sintering data of supported metals can be described numerical simulations presented in this study. Reactor model used in
using n = 1 or 2 for most of the cases. Ψ∞ , residual term, dependent on this work can be found elsewhere [19,20]. Reaction scheme and rate
only aging temperature for a given set of conditions of thermal treat­ expressions used in this work are shown in Table 1
ment, typically, has a negative activation energy Y represents the mole fraction of chemical species, k and Kin repre­
NO oxidation tests in flow reactors revealed that oxidation of Pt to Pt sent Arrhenius reaction rate constant and Inhibition constant, respec­
oxides result in decrease in NO oxidation activity [13,14]. Olsson and tively. Keq represents the equilibrium constant for NO oxidation. Mulla
Fridell [13] confirmed the formation of Pt oxides using XPS studies. et al. [21] elucidated that NO2 inhibits NO oxidation and it has a reac­
They pointed out that metallic Pt is more active for NO oxidation than Pt tion order of -1 when present in excess. Considering such observations,
oxides. Global and semi-micro-kinetic models have been presented to an NO2 based inhibition was used in this work. R2 and R3 have been
capture the impact of PGM oxidation on NO oxidation [15–17]. These used to capture change in PGM oxidation state during NO oxidation.
models incorporated global reactions to capture oxidation of metallic Coverage of metallic PGM, represented by θPt , is incorporated into NO
precious metals to oxides. It is worth pointing out here that all the PGM oxidation rate expression as shown in R1 in Table 1. It should be noted
oxides are lumped into single chemical species (e.g. PtOx) in majority of
the models that incorporated change in PGM oxidation state during NO
oxidation [15–17]. Table 1
The objectives of this work are (i) Develop a hydrothermal aging Reaction scheme and rate expressions.
unified global kinetic model for NO oxidation, (ii) Present a global ki­ Reaction Rate Expression
netic model that captures the change in PGM oxidation state during NO
R1 : NO + 0.5O2 ⇌NO2 1 YNO2
oxidation, and (iii) Extend NO oxidation model at Degreening condition k1 θPt YNO YO2 0.5 (1 −
Keq YNO ∗ YO2 0.5
)
(550 ◦ C-4 h), involving change in PGM oxidation state, to other HTA (1 + Kin YNO 2 )
conditions. R2 : Pt + NO2 →PtOx + NO k2 θPt YNO 2
R3 : PtOx + NO→Pt + NO2 k3 (1 − θPt ) YNO

2
R.K. Dadi et al. Applied Catalysis B: Environmental 283 (2021) 119655

that R2 and R3 are used only in the simulations in which change in oxidation expression shown in R1 of Table 1 was used to simulate bell-shaped
state during NO oxidation is captured. θPt is assumed to be 1 for all the steady-state NO oxidation curves. The model presented in this subsec­
simulations that don’t capture change in PGM oxidation state during NO tion does not capture change in PGM oxidation state during NO oxida­
oxidation. The key reasoning for presenting models with and without tion test. So, coverage of metallic PGM (θPt ) was taken to be unity for all
oxidation state change is to point out the improvement of model’s ac­ the simulations presented in this subsection. Fig. 2 shows bell-shaped
curacy when PGM oxidation is incorporated in the model. curves at HTA temperatures of 550 and 650 ◦ C. A base kinetic model
has been developed at a reference HTA condition, 550 ◦ C–4 h. Kinetic
3.2. Modeling of HTA deactivation of NO oxidation parameters in the rate expression of NO oxidation, estimated by nu­
merical optimization, have been reported in Table 2. Degreening con­
3.2.1. Performance model at discrete HTA conditions dition (550 ◦ C–4 h) was used as the reference condition in the
Fig. 1 shows NO oxidation activity on a Degreened catalyst (HTA: development of the hydrothermal aging model. NO oxidation activity at
550 ◦ C–4 h). As mentioned in experimental section, catalyst is held at a other HTA conditions have been normalized with respect to NO oxida­
given temperature for 35 min. NO oxidation conversion efficiency de­ tion actvities at 550 ◦ C–4 h. Fig. 2 shows steady-state NO oxidation
creases with time, particularly at catalyst temperatures of around 250, curves corresponding to HTA temperatures of 550 ◦ C and 650 ◦ C. NO
300 and 350 ◦ C. Decrease in conversion efficiency is minimal for catalyst oxidation activity decreases from 2 h to 50 h when hydrothermally
temperatures of 150, 400 and 450 ◦ C. Olsson and Fridell [13] illustrated treated at 650 ◦ C. The decrease in NO oxidation activity can be the
the observed decrease in steady-state NO oxidation activity is due to the combined effect of change in metal dispersion and Turn Over Frequency
oxidation of metallic Pt to Pt oxides. Owing to the same reason, i.e. of NO oxidation. The approach presented in this work can only gauge
oxidation of metallic PGM to their oxides, a decrease in NO oxidation overall NO oxidation activity across the catalyst. Metal dispersion data
conversion efficiency is observed. It is apparent from Fig. 1 that NO collected at multiple HTA conditions can be used to deconvolute the loss
oxidation activity does not decrease with time at lower catalyst tem­ in TOF from the change in macroscopic NO oxidation rate [8]. One of the
peratures like 150 ◦ C. Furthermore, the decrease in steady-state activity other limitations of this work is that a more finely resolved data has not
increases with increase in temperature from 200 to 350 ◦ C. This been collected at HTA temperature of 550 ◦ C. NO oxidation activity
behavior can be explained by Arrhenius temperature dependence of could have followed a non-monotonic trend with HTA duration at
PGM oxidation in the presence of oxidizing agents such as O2 and NO2. 550 ◦ C, owing to the non-monotonic trend of NO oxidation activity with
NO oxidation on the catalyst samples investigated here reaches equi­ particle size, particularly in the lower end of particle size (5− 10 nm)
librium regime by around 400 ◦ C and kinetics on precious metal sites do Reactor data sets at all the HTA conditions have been modeled by
not alter conversion efficiency to a considerable extent beyond 400 ◦ C. changing only the pre-exponential factor in the rate expression. A
As a result, catalyst temperatures like 400 and 450 ◦ C resulted in mini­ normalized reaction pre-exponential, which is essentially normalized
mal decrease in NO oxidation activity. A similar reactor tests were reaction rate in this case, has been estimated for each of the HTA con­
performed at other hydrothermal aging conditions to investigate hy­ ditions investigated in this work. Reactor data at HTA: 550 ◦ C–200 h
drothermal stability of NO oxidation on a Diesel Oxidation Catalyst could not be simulated as well as other HTA conditions. Model’s accu­
(DOC). A global kinetic model that does not capture PGM oxidation is racy for this HTA condition could have been improved by changing ki­
shown in this subsection. The raw data shown in Fig. 1 has been aver­ netic parameters other than reaction pre-exponential. However, it has
aged between tend − 60 and tend − 30(s) (tend is the end time of the been observed that the incorporation of PGM oxidation reactions
test-run at each of the catalyst temperatures) to obtain a bell-shaped improved the model’s accuracy at this HTA condition, even without
curve as shown in Fig. 2. Likewise, averaging was performed with changing other kinetic parameters. A more detailed explanation will be
data sets on catalysts samples exposed to different HTA treatments. given in the section, wherein the model on PGM oxidation state is pre­
Although a true steady-state has not been reached at some catalyst sented. Normalized NO oxidation activity was estimated at each of the
temperatures, lower temperatures like 150 or 200 ◦ C do reach nearly HTA conditions by normalizing pre-exponential factors with respect to
steady-state. Consequently, data at these temperatures can be useful Degreening condition. Normalization of rate can simply be expressed by
gauges of NO oxidation intrinsic activity. Even so, we will also present a Eqn. 3. Normalized rates for multiple HTA conditions have been re­
more detailed model that captures change in PGM oxidation state during ported in Table 3. It can be seen from Fig. 1 that NO oxidation activity
NO oxidation reaction also gets affected by the change in PGM oxidation state at catalyst
HTA stability of NO oxidation has been investigated by performing temperature higher than 200 ◦ C. Therefore, it is important to give more
NO oxidation tests under identical set of reactor conditions. Rate weightage to the low temperature points in the evaluation of normalized
pre-exponential factor of NO oxidation intrinsic activity. One of the key
deductions of this work is that modeling the change in PGM oxidation
kinetics improves model prediction of NO oxidation activity in the
catalyst temperature range of 250− 400 ◦ C, particularly under mild HTA
conditions when PGM particle size is low.

rHTA kHTA AHTA exp(− ERTHTA )

= = − E (3)
rref kref Aref exp( RTref )

It was possible to model NO oxidation data under multiple HTA

conditions using identical activation energy in the global rate expression
of NO oxidation (E = 40kJ/mole). Owing to this reason, the ratio of
kinetic constants can be simplified to the ratio of pre-exponential factors
of NO oxidation. The accuracy of the parameters estimated using global
kinetic model is subjected to desired level of accuracy in the prediction
of NO oxidation data. So, a more systematic determination of activation
energies using differential regime data may result in slightly different
values for different HTA conditions. Hauff et al. [22] pointed out that
Fig. 1. NO oxidation experimental data on Degreened Catalyst (HTA: NO oxidation on catalysts with different Pt loading and hydrothermal
550 ◦ C–4 h): Feed: NO = 1000 ppm; O2 = 10 %; 7 % H2O and 8 % CO2 in N2; aging status can be modeled by changing only the active sites,
GHSV = 80,000 h− 1.

3
R.K. Dadi et al. Applied Catalysis B: Environmental 283 (2021) 119655

1
Fig. 2. Model and Data comparison of NO oxidation steady-state curves: Feed: NO = 1000 ppm; O2 = 10 %; 7 % H2O and 8 % CO2 in N2; GHSV = 80,000 h− (a) HTA
Temperature = 550 ◦ C, (b) HTA Temperature = 650 ◦ C.

required to estimate number of active catalytic sites available for

Table 2
oxidation reactions.
Kinetic parameters of NO oxidation: Base HTA (550 ◦ C-4 h).
It can be seen from Fig. 3a that NO oxidation activity monotonically
Rate Constant A Units E Units decreased from 2− 50 h at HTA temperature of 700 ◦ C. NO oxidation
k1 6.4 ∗ 108 /9.2 ∗ 108
[
mol
]
40 (kJ/mole) activity has also been measured at HTA condition of 700 ◦ C-100 h. Data
m .s
3 on catalyst samples with longer duration of HTA treatment
Kin 1.56 [-] − 40 (kJ/mole)
[ ] (550 ◦ C–200 h, 650 ◦ C–50 h, 700 ◦ C–100 h) are critical in the evaluation
k2 1.21 mol 21.3 (kJ/mole) of residual activity at each of the HTA temperatures. The kinetic
[m .s]
3
k3 1.02 ∗ 10 − 2 mol 2.5 (kJ/mole) modeling approach used to develop an HTA deactivation model of NO
m3 .s oxidation is shown in next subsection.

3.2.2. NO oxidation HTA deactivation model

Table 3 General Power Law Expression (GPLE) in Eqn.1 can be analytically
Normalized reaction rates under different HTA conditions. solved to get normalized activity (Ψ) as a function of HTA duration (tage )
HTA specification NO oxidation pre-exp (mol/m3/s) and temperature (Tage ). Eqn. 4 shows HTA time and temperature
550 ◦ C–4 h A = 6.4*108
dependence of normalized activity for first order deactivation (n = 1).
550 ◦ C–200 h 0.9A We present in this subsection how well a first order GPLE model can
650 ◦ C–2 h A describe normalized activities shown in Table 3. We also present
650 ◦ C–20 h 0.7A Arrhenius plots used to estimate parameters shown in Eqn. 4. One of the
650 ◦ C–50 h 0.44A
characteristics of GPLE models is that activity stabilizes at non-zero
700 ◦ C–2 h 0.9A
700 ◦ C–5h 0.7A residual values ,Ψ∞ , after long exposures of thermal treatment. Resid­
700 ◦ C–10h 0.6A ual activity is dependent only on HTA temperature, for a given pre-
700 ◦ C–20 h 0.45A treatment condition. The pre-treatment prior to reactor test can also
700 ◦ C–50 h 0.18A
influence the rate of PGM oxidation during NO oxidation [5,13,14].
700 ◦ C–100 h 0.12A
( )
Ψ = Ψ∞ + (1 − Ψ∞ )exp − kd tage
( ) ( ) (4)
correspondingly reaction pre-exponential in a global kinetic model. E∞ Ed
Ψ∞ = A∞ exp − ; kd = Ad exp −
They also mentioned that direct correlation of catalytic activity with Rg Tage Rg Tage
active sites is only valid if the precious metal particles are not too small
Normalized activity has been estimated after sufficiently long dura­
to be structurally insensitive. It is also important to investigate the
tions of HTA treatments at HTA temperatures of 550 ◦ C (200 h) and
sensitivity of oxidation state of precious metal with HTA. We will pre­
700 ◦ C (100 h). Normalized activities are estimated to be 0.9 and 0.12
sent data to support the above point in the subsequent sections.
after HTA treatments of 550 ◦ C-200 h and 700 ◦ C-100 h, respectively.
Self-inhibition of NO oxidation was modeled using an NO2 based
The reasoning to not include 650 ◦ C data in the calibration of Ψ∞ is that
inhibition as shown in R1 of Table 1. Model captures this self-inhibition
NO oxidation data has not been collected for sufficiently long duration
under HTA condition of 700 ◦ C-50 h, as it can be seen from Fig. 3b. It was
of HTA treatment that can result in the stabilization of activity at 650 ◦ C.
observed that the same set of inhibition parameters predict feed NO
NO oxidation data at HTA T = 650 ◦ C, however, was used to estimate
concentration effect reasonably well for a wide range of HTA conditions
deactivation rate constant. Fig. 4a shows the Arrhenius plot for residual
(HTA: Temperature 550− 1100 ◦ C). This is one of the critical check
activity. 650 ◦ C residual activity has also been included in Fig. 4a.
points to develop an HTA-unified global kinetic model. A systematic
Although, there is no experimental residual data point for this temper­
calibration of self-inhibition of oxidation reactions is paramount to
ature, model predicted value is included in Fig. 4a. It should be noted
develop a HTA unified global kinetic model for oxidation reactions. A
that − ln(Ψ∞ ) versus 1000 has a negative slope, indicating that Ψ∞ has
similar approach can also be used for other oxidation reactions on PGM Tage

based catalysts. It is important to consider the impact of aging on each of negative activation energy and therefore it decreases with increase in
the elementary steps when a more detailed model like micro-kinetic temperature. Lower residual activity at higher HTA temperature is due
model is used. It is likely that only some of the steps in the entire to the increase in particle size with increase in HTA temperature. As
elementary step mechanism get affected by hydrothermal aging. One of mentioned previously, larger particle size results in the lower activity
the other key steps in the development of HTA-unified micro-kinetic once the rate goes past the local maximum with respect to particle size.
model is to get the metal dispersion data as a function of HTA. This is Parameters associated with residual activity are reported in Table 4.
Applying logarithm to Eqn. 4 results in Eqn. 5, which shows linear

4
R.K. Dadi et al. Applied Catalysis B: Environmental 283 (2021) 119655

Fig. 3. Model and Data comparison of NO oxidation steady-state curves at HTA Temperature of 700 ◦ C. O2 = 10 % in Base feed (a) Feed: NO = 1000 ppm; O2 = 10 %;
7 % H2O and 8 % CO2 in N2.
(b) GHSV = 80,000h− 1; HTA Condition: 700 ◦ C–50 h
(c) Feed: NO = 200 ppm; O2 = 10 %; 7 % H2O and 8 % CO2 in N2; GHSV = 60,000 h− 1; HTA Condition: 700 ◦ C–100 h

Fig. 4. (a) Arrhenius plot for residual activity (Ψ∞ ), (b) Deactivation rate constant determination at HTA temperatures of 650 ◦ C and 700 ◦ C.

700 ◦ C (shown in Fig. 4b) suggest that HTA deactivation of NO oxidation

Table 4
can be explained by first order deactivation with sufficient accuracy. kd
Kinetic parameters of the HTA deactivation model of NO oxidation (HTA tem­
values obtained at 650 and 700 ◦ C have been used to estimate pre-
perature < 800 ◦ C).
exponential factor and activation energy of deactivation rate constant
Rate Constant A [− ] E (kJ/mole)
(kd ). Fig. 5a shows Arrhenius plot for kd . It is used to estimate Arrhenius
kd 19.3 114 parameters of deactivation rate constant. It is worth mentioning here
Ψ∞ 2.03∗10− 6 − 89 that 550 ◦ C temperature data was used as validation data of kd , esti­
mated from Fig. 5a. Activation energy of hydrothermal deactivation was
estimated to be 114 kJ/mol. Deactivation energy of sintering was re­
relationship between ln(Ψ − Ψ∞ ) and HTA time (tage ). Fig. 4b shows the ported in a number of articles. However, this is one of the few studies
linearity plots for normalized activities at HTA treatments of 650 and that elucidated a systematic approach to estimate the deactivation en­
700 ◦ C. Slope of these linear plots represent deactivation rate constants ergy of NO oxidation reaction on Pt-Pd based catalyst used in this work.
at the corresponding HTA temperatures. Y-intercepts have been set to −
− ln(Ψ − Ψ∞ ) = − ln(1 − Ψ∞ ) + kd tage (5)
ln(1 − Ψ∞ ) and correlation of Ψ∞ determined from Arrhenius plot
shown in Fig. 4a has been used to calculate the intercepts at 650 and Fig. 5b shows how the Aging model shown in Eqn. 5 capture
700 ◦ C. Linear fits with R2 = 0.99 at HTA temperatures of 650 and

5
R.K. Dadi et al. Applied Catalysis B: Environmental 283 (2021) 119655

Fig. 5. (a) Arrhenius plot for deactivation rate constant, (b) Normalized NO oxidation activity as a function of HTA time and temperature.

normalized activity as a function of HTA duration and temperature. It species on active sites have also been presented [17,23]. Nonetheless, a
can be seen from Fig. 5b that activity decreases precipitously in the first common aspect of majority of these models is that oxidized PGM were
few hours and then the rate of deactivation decreases to finally stabilize assumed to be inactive not only for surface oxidation reaction but also
at the residual activity corresponding to that HTA temperature. The for the adsorption of chemical species. Similarly, rate expression shown
decrease in rate of deactivation can be due to combined effect of stabi­ in R1 of Table 1 was assumed to be linearly proportional to the ratio of
lization in metal dispersion and Turn Over Frequency (TOF) of NO active sites in metallic form. So, it can be implied from this assumption
oxidation. The decrease in rate of loss in TOF has been reported by Liang that all the PGM sites in oxidized form are considered to be inactive for
et al. [8]. They attributed the retardation in loss of TOF to the formation NO oxidation in the model.
of nitrates species on the catalyst surface. HTA stability of NO oxidation Fig. 6a shows the simulation of steady-state NO oxidation experi­
under more severe HTA conditions will be discussed in the next section. mental data corresponding to HTA temperature of 550 ◦ C. The decrease
in steady-state NO oxidation due to oxidation of PGM was reflected in
lower pre-exponential factor of NO oxidation in the global kinetic model
3.3. Modeling of change in PGM oxidation state during NO oxidation that does not capture the change in oxidation state. Getman et al. [24]
pointed out that NO2 dissociation is many orders of magnitude faster
Fig. 1 was used to demonstrate decrease in steady-state NO oxidation than O2 dissociation. Slower dissociation results in slower oxidation of
due to oxidation of metallic PGM to their oxides. A modeling approach precious metals. Therefore, ratio of pre-exponentials of PGM oxidation
to simulate this behavior is presented in this subsection. In addition to in presence of NO2 and O2 is typically very high (~10,000) [16]. Owing
NO oxidation reaction (R1), R2 and 3 reactions (Table 1) were used to to this reason, PGM oxidation in presence of O2 is not considered in this
capture time on stream deactivation of NO oxidation. R2 represents work. This approach captures NO oxidation data under Degreened
oxidation reaction of PGM in the presence of NO2. R3 represents condition (HTA: 550 ◦ C–4 h), shown in Fig. 6a, reasonably well. Kinetic
reduction of PGM oxides in presence of NO. R2 and 3 were used in the parameters corresponding to PGM oxidation were reported in Table 2.
reactor model to numerically compute θPt , which is the proportion of Normalized reaction rates of NO oxidation at different HTA condi­
active sites that are in metallic form. Fraction of active PGM sites in tions have been tabulated in Table 3. An aging model has also been
oxidized state was represented by 1 − θPt . A notable point here is that developed based on this table. A key point to be noted is catalyst tem­
different kinds of PGM oxides that can potentially be formed when peratures of 150 ◦ C and 200 ◦ C are not so high that they result in sig­
exposed to oxidizing agents like NO2 or O2 are lumped into single nificant decrease in NO oxidation activity due to the change in PGM
oxidized species and their ratio to total active sites was represented as oxidation state. So, the model that does not capture change in oxidation
1 − θPt . This modeling approach, which can capture change in PGM state during NO oxidation is still useful in evaluating the intrinsic NO
oxidation state during NO oxidation has been reported by Arvajova and oxidation activity at low temperatures like 150 or 200 ◦ C. A model that
coworkers [15,16]. We deployed a similar global kinetic modeling captures change in PGM oxidation state improves the model accuracy
approach in this work. More detailed models that involve adsorption of

Fig. 6. Experimental Data and Model prediction of steady-state decrease in NO conversion. (a) HTA Temperature = 550 ◦ C, (b) HTA Temperature = 650 ◦ C. Feed:
NO = 1000 ppm; O2 = 10 %; 7 % H2O and 8 % CO2 in N2.

6
R.K. Dadi et al. Applied Catalysis B: Environmental 283 (2021) 119655

when catalyst temperature is in the range of 250− 350 ◦ C. In accordance

with the above statements, the same correlation for pre-exponential
factor of NO oxidation developed in the previous section has been
used to model the transient data that involves a steady drop in NO
oxidation activity. Model prediction of transient NO oxidation data
corresponding to HTA condition of 550 ◦ C–200 h has been shown in
Fig. 6a. In addition to the pre-exponential factor of NO oxidation, the
pre-exponential factor of PGM oxidation has also been tuned to model
the experimental data at multiple HTA conditions. All the kinetic pa­
rameters but the pre-exponential factors of reactions: R1 and R2 were
kept unchanged with change in HTA conditions. Pre-exponential factors
of R2 for a few HTA conditions have been tabulated in Table 5. Fig. 6b
shows model prediction of NO oxidation activity corresponding to HTA
temperature of 650 ◦ C. Fig. 6b can be used to emphasize the applica­
bility of this modeling approach to different HTA temperatures. The
change in reaction pre-exponential factor of R2 with HTA, as shown in
Table 5, points out the sensitivity of PGM oxidation rate with HTA Fig. 7. Experimental Data and Model prediction of steady-state decrease in NO
treatment, consequently with particle size. Olsson and Fridell [13] conversion. Feed: NO = 1000 ppm; O2 = 10 %; 7 % H2O and 8 % CO2 in N2;
concluded that smaller Pt particles form oxides more easily compared to GHSV = 80,000 h− 1.
larger particles. Fig. 6 depicts the time on stream deactivation of NO
oxidation at hydro-thermal aging temperatures of 550 ◦ C and 650 ◦ C. 1100 ◦ C) can be seen from Fig. 8. HTA model with aging parameters
The impact of hydrothermal aging duration on NO oxidation at HTA shown in Table 4 is not able to accurately capture the model predicted
temperature of 550 ◦ C can predominantly be captured by tuning only the normalized activities under extreme HTA conditions (HTA temperature
kinetic parameters of PGM oxidation. Fig. 2 shows that difference in NO ≥ 800 ◦ C). This is due to the well-known phenomenon of change in
oxidation activities at low catalyst temperatures like 150 or 200 ◦ C is sintering mechanism of PGM particles beyond a particular HTA tem­
almost insignificant between 550 ◦ C–4 h and 550 ◦ C–200 h. However, a perature, typically in the range of 700− 800 ◦ C [25]. A mechanistic HTA
considerable difference was observed in NO oxidation activities at deactivation model [26] would have been able to capture the normal­
catalyst temperatures between 250− 400 ◦ C. Model that includes ized activities over a wide range of HTA temperatures. One of the lim­
oxidation of PGM captures the aforementioned behavior. From a itations of GPLE or DMRA models is that they have to be calibrated
mechanistic perspective, NO oxidation rate is mainly influenced by the differently under extreme HTA conditions [26]. Similarly, in this work,
PGM oxidation rate at mild hydro-thermal aging temperatures like deactivation model was calibrated differently under extreme HTA con­
550 ◦ C. It has been demonstrated using Fig. 6 that time on stream ditions. Kinetic parameters associated with extreme HTA deactivation
deactivation of NO oxidation occurs at HTA temperatures of both 550 ◦ C model are tabulated in Table 6.
and 650 ◦ C. This time on stream deactivation, primarily led by PGM The authors would like to point out an observation on the difference
oxidation, is dependent on HTA temperature. Fig. 7 shows that time on in HTA deactivation parameters shown in Tables 4 and 6. The aging
stream deactivation of NO oxidation is insignificant when the catalyst model under extreme HTA conditions has a higher kd and Ψ∞ as
was exposed to HTA treatment of 700 ◦ C–50 h. A global kinetic model compared to the moderate HTA model (at some reference HTA tem­
without PGM oxidation and reduction reactions can describe data in perature like 800 ◦ C). A higher kd indicates a steeper loss in the activity
Fig. 7 with sufficient level of accuracy. This suggests that PGM oxidation and a higher Ψ∞ indicates a higher residual activity. These aspects can
rate, consequently, time on stream deactivation of NO oxidation, gets also be intuitively explained from an atomic level sintering mechanism.
significantly attenuated when PGM particle becomes sufficiently large, An extreme HTA results in a steeper increase in the PGM particle size,
e.g. particle size corresponding to HTA treatment of 700 ◦ C–50 h. It consequently, a steeper initial decrease in activity was observed. A
should be noted that catalyst inlet temperature profile used for 700 ◦ C higher Ψ∞ is perhaps representative of the slowdown of particle
has some differences as compared to 550 and 650 ◦ C. Although not agglomeration due to the increase in interparticle diffusional limitations
shown here, time on stream deactivation is also observed to be very at higher HTA temperatures. The rate of decrease in residual NO
minimal for other aging durations (5,10, 20 h) at HTA Temperature
=700 ◦ C.

3.4. Additional modeling discussion

3.4.1. HTA deactivation model beyond 700 ◦ C

Diesel engine exhaust temperatures on different applications across
the globe can be up to 800 ◦ C. Therefore, it is important to investigate
hydrothermal stability of NO oxidation beyond 700 ◦ C. HTA stability of
NO oxidation under extreme HTA conditions (up to HTA temperatures of

Table 5
Reaction pre-exponential factors of PGM oxidation.
HTA Reaction pre-exponential factor of PGM oxidation reaction
specification (mol/m3/s)

550 ◦ C–4 h A2
550 ◦ C–200 h 2.11 A2
650 ◦ C–2 h 2.11 A2 Fig. 8. Experimental Data and Model prediction of steady-state NO oxidation
650 ◦ C–20 h 1.65 A2
under extreme HTA conditions. Feed: NO = 200 ppm; O2 = 10 %; 7 % H2O and
650 ◦ C–50 h A2
8 % CO2 in N2; GHSV = 60,000 h− 1.

7
R.K. Dadi et al. Applied Catalysis B: Environmental 283 (2021) 119655

Table 6 550 ◦ C. The activity drop in NO oxidation rate is mainly by PGM

Kinetic parameters of the HTA deactivation model of NO oxidation (HTA tem­ oxidation at mild aging temperatures like 550 ◦ C. So, the activity on
perature ≥ 800 ◦ C). metallic sites is not expected to decrease considerably at HTA Temper­
Rate Constant A [− ] E (kJ/mole) ature = 550 ◦ C. PGM oxidation model captures the deactivation due to
formation of PGM oxides. The convolution of deactivation in NO
kd 110 114
oxidation rate per unit active site and decrease in number of active sites
9.27∗10− 4 − 41
can be one of the plausible reasons for the second order deactivation
Ψ∞

model to not predict expected trend at HTA temperature = 550 ◦ C

oxidation activity with respect to HTA temperature was observed to
3.4.3. Application of hydro-thermal aging model
decrease under extreme HTA conditions.
One of the most important practical applications of NO oxidation
deactivation model is in automotive emission control. Catalyst temper­
3.4.2. Comparison of deactivation models with n = 1 and n = 2
ature, therefore aging temperature is a transient function of time
Sintering data in the literature could be fitted to either first order or
(Tage (t) = f(t)). Eqn. 1 with n = 1 can be used to calculate normalized
second order deactivation with a sufficient level accuracy [11]. Some of
the fundamental processes associated with atomic level sintering activity as a function of time. Eqn. 1 is an ordinary differential equation
mechanism involve second order processes. So, it is worthwhile to with temperature dependent aging parameters. This ODE needs to be
investigate if the normalized activities can be described with second solved numerically since aging temperature is typically a dynamic
order deactivation model. Fig. 9a shows normalized activity curves function of time under real drive cycle operations. Runge Kutta ODE
when second order deactivation model (Ed = 200 kJ/mol) was used. The solvers can be used to numerically solve ODE in Eqn. 1. A key point to be
second order deactivation model was able to describe normalized ac­ highlighted here is that the deactivation parameters presented for
tivity curves qualitatively well until aging time = 200 h. However, the extreme HTA conditions should be used in Eqn.1 when aging tempera­
error in second order model prediction of normalized activity is higher ture is higher than 800 ◦ C.
than n = 1. Furthermore, second order deactivation model takes much
longer duration to reach asymptote at HTA Temperature = 550 ◦ C as 3.4.4. Modeling approach for other oxidation reactions
compared to first order deactivation model. It is more appropriate to use This paper demonstrates a modeling approach to develop HTA model
first order deactivation than to use second order since activity is ex­ for NO oxidation on PGM catalysts. A similar approach can be used for
pected to reach asymptote by 500 h even at mild HTA temperatures like other oxidation reactions like CO or HC oxidation reactions. A critical
step is to collect reactor data at a wide range of HTA conditions. A base
kinetic model that accounts for all the key mechanistic phenomena
should be developed. Subsequently, the base kinetic model needs to be
extended to other HTA conditions. Model estimated normalized activ­
ities should be used to develop deactivation models. CO oxidation can be
used as an example to further demonstrate this approach. Langmuir-
Hinshelwood kinetics model can be developed for a base HTA condi­
tion. CO oxidation data at other HTA conditions can be captured by
tuning only the reaction pre-exponential in LH rate expression. A similar
approach as demonstrated above can be used to develop an HTA model
for CO oxidation once the global kinetic model is calibrated for a wide
range of HTA conditions.

4. Summary and conclusions

We developed a hydrothermal aging model of NO oxidation for a

state-of-the-art Diesel Oxidation Catalyst. This aging model has been
developed based on the flow-reactor data of NO oxidation. A systematic
modeling approach has been presented to translate the reactor data at
each of the HTA conditions to the normalized reaction rates. Normalized
reaction rates have been used to develop a first order deactivation model
(Ed = 114kJ/mol) with residual activity. This paper presents the
following novel experimental and modeling results on Pt-Pd oxidation
catalyst:

1 NO oxidation reactor data on a Pt-Pd catalyst has been presented for

a wide range of HTA temperatures (550 ◦ C–1100 ◦ C). To the best
knowledge of authors, there does not exist a previous study that
covered as wide range of HTA conditions as presented in this work
for NO oxidation.
2 A systematic approach was presented to develop an HTA deactiva­
tion model of NO oxidation. A practical application of this aging
model under real drive cycle operations was also presented. We also
highlighted the necessity of recalibration in deactivation model pa­
rameters under extreme HTA conditions when empirical aging
models like GPLE models are used. A number of articles reported this
Fig. 9. (a) Normalized NO oxidation activity as a function of HTA time and phenomenon in the past for sintering models on PGM based catalysts.
temperature with n = 2 (order of deactivation). (b) Normalized activity curves We elucidated this phenomenon for NO oxidation quantitatively.
at HTA Temperature = 550 ◦ C with n = 1 and 2.

8
R.K. Dadi et al. Applied Catalysis B: Environmental 283 (2021) 119655

3 The experimental and modeling work presented in this work [2] T.R. Johns, R.S. Goeke, V. Ashbacher, P.C. Thüne, J.W. Niemantsverdriet, B. Kiefer,
A.K. Datye, Relating adatom emission to improved durability of Pt–Pd diesel
demonstrate the impact of HTA treatment on PGM oxidation,
oxidation catalysts, J. Catal. 328 (2015) 151–164.
consequently on, time on stream deactivation of NO oxidation. A [3] S.B. Kang, M. Hazlett, V. Balakotaiah, C. Kalamaras, W. Epling, Effect of Pt: Pd ratio
number of articles in the past showed the modeling approach to on CO and hydrocarbon oxidation, Appl. Catal. B: Environ. 223 (2018) 67–75.
capture the change in PGM oxidation state during NO oxidation at a [4] T.C. Watling, M. Ahmadinejad, M. Ţuţuianu, Å. Johansson, M.A. Paterson,
Development and validation of a Pt-Pd diesel oxidation catalyst model, Sae Int. J.
particular HTA condition. One of the unique elements of this work is Engines 5 (2012), 01-1286.
that we presented a modeling approach to capture the impact of HTA [5] B.M. Shakya, B. Sukumar, Y.M. López-De Jesús, P. Markatou, The effect of Pt: Pd
treatment on time on stream deactivation of NO oxidation. Most ratio on heavy-duty diesel oxidation catalyst performance: an experimental and
modeling study, SAE Int. J. Engines 8 (2015), 01-1052.
notably, we highlighted the impact of particle size on PGM oxidation [6] G.W. Graham, H.W. Jen, O. Ezekoye, R.J. Kudla, W. Chun, X.Q. Pan, R.W. McCabe,
at mild HTA temperatures like 550 ◦ C. Modeling study in this work Effect of alloy composition on dispersion stability and catalytic activity for NO
confirmed the change in PGM oxidation rate is the main contributor oxidation over alumina-supported Pt–Pd catalysts, Catal. Letters 116 (1–2) (2007)
1–8.
to the change in NO oxidation rate with respect to aging duration at [7] J. Kim, C. Kim, S.J. Choung, Comparison studies on sintering phenomenon of diesel
mild HTA temperatures like 550 ◦ C. Data presented in this work also oxidation catalyst depending upon aging conditions, Catal. Today 185 (1) (2012)
emphasizes that time on stream deactivation of NO oxidation de­ 296–301.
[8] Y. Liang, C. Ou, H. Zhang, X. Ding, M. Zhao, J. Wang, Y. Chen, Advanced insight
creases significantly when HTA temperatures (≥ 700 ◦ C) into the size effect of PtPd nanoparticles on NO oxidation by in situ FTIR spectra,
4 A systematic approach was presented to develop HTA unified global Ind. Eng. Chem. Res. 57 (11) (2018) 3887–3897.
kinetic models for oxidation reactions. HTA unified global kinetic [9] A. Monzón, T.F. Garetto, A. Borgna, Sintering and redispersion of Pt/γ-Al2O3
catalysts: a kinetic model, Appl. Catal. A: Gen. 248 (1–2) (2003) 279–289.
model was presented for NO oxidation in this work. We illustrated
[10] W. Arabczyk, I. Jasińska, Z. Lendzion-Bieluń, Kinetics studies of recrystallization
the approach that can be used for other oxidation reactions like CO process of metallic catalysts for ammonia synthesis, Catal. Today 169 (1) (2011)
and HC oxidation. 93–96.
5 A number of previous articles presented models that predict either [11] C.H. Bartholomew, Sintering kinetics of supported metals: new perspectives from a
unifying GPLE treatment, Appl. Catal. A: Gen. 107 (1) (1993) 1–57.
the effect of aging or the effect of PGM oxidation on NO oxidation. [12] A. Monzón, E. Romeo, A. Borgna, Relationship between the kinetic parameters of
The approach used in this work captures the impact of both PGM different catalyst deactivation models, Chem. Eng. J. 94 (1) (2003) 19–28.
oxidation and hydrothermal aging on NO oxidation. NO oxidation [13] L. Olsson, E. Fridell, The influence of Pt oxide formation and Pt dispersion on the
reactions NO2⇔ NO+ 1/2 O2 over Pt/Al2O3 and Pt/BaO/Al2O3, J. Catal. 210 (2)
activity at low temperatures like 150 or 200 ◦ C are used to estimate (2002) 340–353.
activity on metal sites. Time on stream deactivation of NO oxidation [14] K. Hauff, U. Tuttlies, G. Eigenberger, U. Nieken, Platinum oxide formation and
activity, which predominantly happens in the catalyst temperature reduction during NO oxidation on a diesel oxidation catalyst–experimental results,
Appl. Catal. B: Environ. 123 (2012) 107–116.
range of 250− 400 ◦ C, is used to quantify the HTA impact on PGM [15] A.B. Arvajová, P. Boutikos, R. Pečinka, P. Kočí, Global kinetic model of NO
oxidation during NO oxidation. oxidation on Pd/γ-Al2O3 catalyst including PdOx formation and reduction by CO
and C3H6, Appl. Catal. B: Environ. 260 (2020), 118141.
[16] A. Arvajová, P. Kočí, V. Schmeißer, M. Weibel, The impact of CO and C3H6 pulses
CRediT authorship contribution statement on PtOx reduction and NO oxidation in a diesel oxidation catalyst, Appl. Catal. B:
Environ. 181 (2016) 644–650.
Rama Krishna Dadi: Conceptualization, Methodology, Software. [17] W. Hauptmann, M. Votsmeier, J. Gieshoff, A. Drochner, H. Vogel, Inverse
hysteresis during the NO oxidation on Pt under lean conditions, Appl. Catal. B:
Rohil Daya: Conceptualization, Methodology, Software. Ashok
Environ. 93 (1–2) (2009) 22–29.
Kumar: Investigation. Saurabh Y. Joshi: Conceptualization, Method­ [18] A. Kumar, M.A. Smith, K. Kamasamudram, N.W. Currier, A. Yezerets, Chemical
ology, Software. Hongmei An: Investigation. Michael J. Cunningham: deSOx: an effective way to recover Cu-zeolite SCR catalysts from sulfur poisoning,
Supervision. Neal W. Currier: Supervision. Aleksey Yezerets: Catal. Today 267 (2016) 10–16.
[19] AVL BOOST Aftertreatment Manual, AVL, (2019). http://www.avl.com.
Supervision. [20] R. Daya, S.Y. Joshi, J. Luo, R.K. Dadi, N.W. Currier, A. Yezerets, On kinetic
modeling of change in active sites upon hydrothermal aging of Cu-SSZ-13, Appl.
Declaration of Competing Interest Catal. B: Environ. 263 (2020), 118368.
[21] S.S. Mulla, N. Chen, W.N. Delgass, W.S. Epling, F.H. Ribeiro, NO 2 inhibits the
catalytic reaction of NO and O 2 over Pt, Catal. Letters 100 (3–4) (2005) 267–270.
The authors declare that they have no known competing financial [22] K. Hauff, U. Tuttlies, G. Eigenberger, U. Nieken, A global description of DOC
interests or personal relationships that could have appeared to influence kinetics for catalysts with different platinum loadings and aging status, Appl. Catal.
B: Environ. 100 (1–2) (2010) 10–18.
the work reported in this paper. [23] R. Khazanchi, D. Reddy, D. Bhatia, Kinetic model for the reversible deactivation of
a Pt/Al2O3 catalyst during NO oxidation, Chem. Eng. J. 314 (2017) 139–151.
Acknowledgements [24] R.B. Getman, W.F. Schneider, A.D. Smeltz, W.N. Delgass, F.H. Ribeiro, Oxygen-
coverage effects on molecular dissociations at a Pt metal surface, Phys. Rev. Lett.
102 (7) (2009), 076101.
The authors would also like to thank Brian Mnichowicz for his [25] D.D. Beck, C.J. Carr, A study of thermal aging of PtAl2O3 using temperature-
contribution to this work. programmed desorption spectroscopy, J. Catal. 110 (2) (1988) 285–297.
[26] E.E. Finney, R.G. Finke, Catalyst sintering kinetics data: is there a minimal
chemical mechanism underlying kinetics previously fit by empirical power-law
References expressions and if so, what are its implications? Ind. Eng. Chem. Res. 56 (37)
(2017) 10271–10286.
[1] A. Morlang, U. Neuhausen, K.V. Klementiev, F.W. Schütze, G. Miehe, H. Fuess, E.
S. Lox, Bimetallic Pt/Pd diesel oxidation catalysts: structural characterisation and
catalytic behaviour, Appl. Catal. B: Environ. 60 (3–4) (2005) 191–199.

View publication stats