Mechanistic effects of Water on Carbon monoxide and propylene oxidation on platinum and palladium bimetallic catalysts

Journal Pre-proof Mechanistic Effects of Water on Carbon Monoxide and Propylene Oxidation on Platinum and Palladium Bimetallic Catalysts Melanie J. Hazlett (Conceptualization) (Data curation) (Formal analysis) (Investigation) (Methodology) (Visualization) (Writing original draft), William S. Epling (Formal analysis) (Funding acquisition) (Project administration) (Supervision) (Writing - review and editing)

PII:

S0920-5861(20)30023-7

DOI:

https://doi.org/10.1016/j.cattod.2020.01.024

Reference:

CATTOD 12643

To appear in:

Catalysis Today

Received Date:

13 September 2019

Revised Date:

16 December 2019

Accepted Date:

19 January 2020

Please cite this article as: Hazlett MJ, Epling WS, Mechanistic Effects of Water on Carbon Monoxide and Propylene Oxidation on Platinum and Palladium Bimetallic Catalysts, Catalysis Today (2020), doi: https://doi.org/10.1016/j.cattod.2020.01.024

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2020 Published by Elsevier.

Mechanistic Effects of Water on Carbon Monoxide and Propylene Oxidation on Platinum and Palladium Bimetallic Catalysts

of

Melanie J. Hazlett1, and William S. Epling2*

1

Department of Chemical and Biomolecular Engineering, University of Houston, Houston TX Department of Chemical Engineering, University of Virginia, Charlottesville, VA *[email protected]

-p

ro

2

Jo

ur na

lP

re

Graphical abstract

1

Highlights   

ro

of



Effects of water on CO and C3H6 oxidation were studied over Pt, Pd catalysts and bimetallic Pt-Pd catalysts. Water improved CO oxidation light-off for Pd and Pt-Pd catalysts and had a negative impact on Pt. More CO adsorbed on Pd and Pt-Pd catalysts when water was present, particularly at the particle/support interface. For C3H6 oxidation, the Pt-Pd catalyst showed intermediate light-off behavior between the two monometallic catalysts.

ABSTRACT

-p

Low temperature combustion (LTC) diesel engines are more fuel efficient and have coincident

re

lower temperature exhaust gas and higher CO and hydrocarbon exhaust gas concentrations than today’s standard diesel engine. To meet regulations, diesel oxidation catalysts (DOCs) will need

lP

to be improved and optimized to handle these higher CO and hydrocarbon concentrations emitted and the simultaneous lower exhaust temperatures. Pt-Pd bimetallic catalysts are often used as

ur na

oxidation catalysts. Here, the mechanistic effects of water on CO and C3H6 oxidation were studied over model monometallic Pt and Pd catalysts, and a bimetallic 1:1 Pt-Pd/ γ-Al2O3 catalyst. Water in the reaction mixture improved CO oxidation light-off for the monometallic Pd and Pt-Pd catalysts and had a negative impact on light-off over the monometallic Pt catalyst. Diffuse

Jo

reflectance infrared Fourier transform spectroscopy (DRIFTS) showed that more CO adsorbed on the Pd and Pt-Pd catalysts when water was present, particularly at the particle/support interface, while the opposite occurred on the Pt catalyst. For C3H6 oxidation, water in the reaction showed a negative impact on the light-off temperature for the Pd catalyst and a positive effect for the Pt catalyst. The Pt-Pd catalyst showed intermediate light-off behavior between the two monometallic

2

catalysts. In terms of selectivity, slightly more CO and acetic acid were formed over the Pd and Pt-Pd catalysts when water was in the reaction mixture, while less CO was formed over the Pt catalyst, but overall, the selectivity towards certain partial oxidation products was not greatly affected with the addition of water to the reaction mixture.

KEYWORDS

of

Oxidation catalyst; CO oxidation; Propylene oxidation; Bimetallic Pt:Pd catalysts; Oxidation

ro

mechanism

-p

INTRODUCTION

The role of water in CO and hydrocarbon oxidation mechanisms has implications in a wide range

re

of applications, from automotive exhaust control to proton exchanged membrane (PEM) fuel cells.

lP

For instance, for natural gas vehicle exhaust, where the exhaust contains large water concentrations from methane combustion, the role of water in the oxidation of methane over Pt and Pd, and PtPd/Al2O3 catalysts has been studied and was found to have an inhibiting effect on the oxidation

ur na

rate,1,2 and can lead to catalyst deactivation.3 Moreover, hydrothermal treatments on Pt-Pd/Al2O3 catalysts may reverse Pd surface segregation that can occur over the catalyst lifetime.1 For gasoline engine exhaust, the effect of water on a three-way catalyst has been evaluated and some

Jo

enhancement of CO and propylene oxidation over Pd catalysts was found.4 For diesel emissions, studies have shown that adding water to the reaction feed enhances low

temperature CO oxidation on Pt/Al2O3 catalysts5 and both CO oxidation and propylene oxidation over Pd/Al2O3 catalysts.6 Over a Pd/Al2O3 catalyst, isotopic labeling experiments demonstrate that low temperature CO and propylene oxidation mechanisms involve water as the oxidant, with oxygen playing a secondary role.6 This is consistent with findings from CO oxidation on Pt and 3

Pd catalysts for preferential CO oxidation (PROX) for PEM fuel cells. On a Pd/CeO2-TiO2 catalyst, low temperature CO oxidation was attributed to reactions involving OH from the water; water also suppresses some carbonate species formation on the surface which leads to better CO oxidation.7 On Al2O3-supported Pt and Pd catalysts, infrared characterization studies revealed that dispersed metal cations as well as metal oxide species were responsible for changing the behavior of surface aluminum OH sites and Lewis acid sites.8 These sites were regenerated via re-oxidation

of

with oxygen and water at moderate temperatures. Microkinetic modeling of CO oxidation on Pt

ro

suggests carboxyl intermediated CO oxidation in the presence of water.9 Without water present, the mechanism for CO oxidation on Pt-based catalysts has been stipulated to be carbonate

-p

intermediated.10 However, in a recent study, the authors concluded that formation of surface

re

carbonates acted as a poison, reducing the activity of Pd and Pd-rich Pt-Pd bimetallic catalysts.11 During co-oxidation of CO and propylene, propylene oxidation is limited until the CO is reacted

lP

due to strong CO interaction with the precious metal sites.

There has been increasing momentum towards using higher efficiency lean combustion modes

ur na

in the automobile industry for fuel economy gains. The emissions from these lean combustion modes, or low temperature combustion technologies, differ from current engine platforms and therefore changes in catalytic aftertreatment systems may be required. These lean combustion fuel modes, using reactivity controlled combustion ignition (RCCI) as an example, emit much higher

Jo

CO and hydrocarbon emissions than a conventional diesel engine mode and have a lower exhaust temperature.1 These higher CO and hydrocarbon emissions are also present for other advanced compression ignition (ACI) combustion modes, such as homogeneous charge compression ignition (HCCI).13 ACI combustion modes are being studied for the U.S. Department of Energy’s Co-Optimization of Fuels & Engines (Co-Optima) initiative, where the fuel being investigated is

4

a mixture of conventional liquid fuels and blendstocks produced from various resources including non-food biomass sources.14,15 This combination highlights a need to develop oxidation catalysts with higher activity. Bimetallic Pt-Pd catalysts have received a lot of attention due to their relatively high oxidation activity, which can change as a function of Pt:Pd ratio,16 however the effect of water on these

of

bimetallic catalysts in oxidizing CO and propylene is not as widely studied as on their monometallic counterparts. In order to evaluate the monometallic and bimetallic Pt-Pd catalysts

ro

for diesel oxidation catalysts (DOCs) on low temperature combustion engines, the effect of water

-p

on these oxidation reactions is studied. Reaction light off is important for these applications to reduce cold start emissions, the lower the light off temperature the better. However, catalyst

re

screening often occurs by using a probe reaction, such as CO oxidation, without including other reactants which are ubiquitous in exhaust conditions and often assumed to have a negligible or

lP

consistent effect on catalyst performance. Moreover, many studies in the literature for vehicle exhaust catalyst applications continue to use pretreatment conditions with oxygen only in an inert

ur na

carrier (without H2O present), or reaction conditions without water being present. This can be problematic for catalyst screening and reporting results in literature when the conditions vary from industrial practice, which do use H2O in the pretreatment and testing protocols.17 In terms of the catalytic reaction, the presence of water can either inhibit or enhance oxidation of CO and

Jo

hydrocarbons on catalyst surfaces by competitive adsorption or the water acting as an oxidant itself, as was noted in the discussion above. In terms of the pretreatment, water and oxygen both can oxidize the catalyst surface, which impacts catalytic performance. As water it ubiquitous in automotive exhaust, it is important to evaluate the impact of water on both the catalyst and catalyst support in pretreatment and reaction conditions. 5

In the present study we investigated the effect of water on a 1:1 Pt-Pd/Al2O3 catalyst and the two monometallic Pt and Pd catalysts. The effect of water on CO oxidation and propylene oxidation, individually, was measured and surface species on Pt, Pd, and Pt-Pd/Al2O3 catalysts were probed using diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS). Adding water in the pretreatment and reaction conditions led to sometimes opposite results between the three catalysts and changed the overall reaction pathway, demonstrating that performing catalyst screening

ro

of

experiments under more relevant reaction conditions is critical.

EXPERIMENTAL METHODS

-p

Catalyst Synthesis. Honeycomb cordierite monoliths washcoated with γ-Al2O3 were supplied by Johnson Matthey, with a 1.59 g/in3 Al2O3 washcoat loading and 400 cells/in2. Three catalysts

re

were made using incipient wetness impregnation; Pd, 1:1 Pt:Pd, and Pt. The platinum group metal

lP

(PGM) precursors were Pd(NO3)2 and Pt(NH3)4(NO3)2, which were both purchased from Sigma Aldrich. These were weighed and mixed into a deionized water solution such that the desired metal loading would fill the alumina pore volume. The loading of the catalysts made were 0.55

ur na

wt% PGM for the Pd/Al2O3 catalyst, 0.78 wt% PGM for the 1:1 Pt:Pd/Al2O3 catalyst, and 1 wt% PGM for the Pt/Al2O3 catalyst. These loadings were selected such that the number of moles of PGM on each catalyst were identical. The catalysts were then dried and calcined in a Neytech

Jo

Vulcan 3-550 muffle furnace for 4 hours at 550°C, then aged at 700°C in situ under flowing 14% O2, 5% H2O in balance N2 for 24 hours. The particle sizes as measured by CO pulse chemisorption experiments for these catalysts were 19.4 nm, 5.2 nm, and 4.3 nm for the Pt, 1:1 Pt:Pd, and Pd catalysts respectively. This corresponds to a dispersion of 5.8%, 21.7%, and 26.2% respectively. Details for these pulse chemisorption experiments are presented in previous work18.

6

Reactor Experiments. The catalysts were cut down to 2 inches in length and an appropriate inner diameter to fit into the reactor. The catalysts were placed in a quartz tube which was 1 inch in inner diameter and were wrapped in insulation to avoid gas bypass. Small, hollow quartz tubes were placed upstream to avoid fully developed flow patterns. The gas flow rate used was approximately 11.5 L/min, to maintain a 50,000 hr-1 gas hourly space velocity for each experiment. The gas flow was controlled by Bronkhorst mass flow controllers, and the water was evaporated

of

and flow rate controlled by a Bronkhorst Controlled Evaporator Mixer system.

ro

Temperature programmed oxidation (TPO) experiments were performed with a 7.3°C/min ramp rate of the gas measured upstream of the catalyst, from 80 to 300°C. Upstream of the reactor, the

-p

inlet gas stream was heated by a preheater that was ramped in temperature during the experiment

re

in order to keep the temperature gradient along the catalyst length less than 3°C, as measured during an experiment with N2 only flowing. The outlet gases from each TPO were measured with

lP

an MKS MultiGas 2030 FTIR gas analyzer. Based on a repeat experiment the standard deviation of these TPO experiments is 0.4°C.

ur na

The gas concentrations during the TPO experiments were selected in order to be representative of the high hydrocarbon and CO concentrations observed in the exhaust from low temperature combustion engines. The concentration levels tested were as follows. 

For CO oxidation experiments the concentrations in the inlet gas were 3000 ppm CO,

Jo

10% O2, in balance nitrogen with either 0 or 5% H2O.



For a hydrocarbon, propylene was selected, and the concentrations in the inlet gas were 1500 ppm C3H6, 10% O2 in balance nitrogen with either 0 or 5% H2O.

In the results presented, the data obtained with 0 and 5% H2O during the TPO experiments are labeled as dry and wet. Before each TPO experiment, a pretreatment was performed, lasting 1

7

hour at 600°C. For each 0% H2O and 5% H2O experiment, a dry or wet pretreatment was conducted with the following concentrations; 14% O2 and 0 or 5% H2O in balance N2. DRIFTS Experiments. In order to better understand the CO oxidation mechanism over each catalyst, in situ DRIFTS experiments were performed using a Nicolet 6700 spectrometer equipped with an MCT detector and a high temperature Harrick Scientific Praying Mantis reaction chamber with ZnSe windows. The washcoat was scraped off the cordierite honeycomb monolith and 30

of

mg was pressed into a pellet and placed into the reaction chamber. The DRIFTS spectra were

ro

collected in the 4000 – 650 cm-1 wavenumber range, accumulating 98 scans at 4 cm-1 resolution. The reflectance spectrum was converted to KM units in the OMNIC software.

-p

TPO experiments were performed to mirror the conditions used for the reactor experiments,

re

except the ramp rate of the gas stream was slowed to 4.2°C/min in order to increase the number of scans acquired in a smaller temperature interval during each experiment, and the temperature ramp

lP

was from 66-300°C. A background spectrum was taken at the beginning of the temperature ramp in flowing He, and then the reactant gases were added and the samples were exposed to the feed

ur na

gas for at least 1 hour before the temperature ramp was started. The temperature used in the following figures is the outlet gas stream temperature obtained from flowing 50 mL/min He during the same temperature ramp procedure as the TPO experiments. The concentrations during the DRIFTS TPO experiments were identical to those used in the

Jo

reactor testing; 3000 ppm CO, with 10% O2 and 0 or 5% H2O in balance He. A 50 mL/min total flow rate was maintained using MKS mass flow controllers. For water addition, a Bronkhorst Controlled Evaporator Mixer system was again used. TPO experiments were also performed with O2 and He (with and without H2O) only so the spectral data obtained could be subtracted from the spectra obtained during the TPOs with CO. This allowed removal of the background shift caused

8

by temperature and gas-phase water features.

An analogous pretreatment to the reactor

experiments was used before the experiments and between each TPO experiment - 550°C with 14% O2 in N2 with or without H2O for 1 hour respectively for the dry and wet experiments.

RESULTS AND DISCUSSION CO Oxidation. CO oxidation over the Pd/Al2O3 catalyst was impacted by the presence of water

of

in the reaction and pretreatment mixtures, shown in Figure 1. The temperature to reach 50% conversion (T50) under dry conditions was 146 °C, while under wet conditions the T50 decreased

ro

to 136 °C. In contrast, for the Pt/Al2O3 catalyst, shown in Figure 1, water in the reaction mixture

-p

inhibited CO oxidation, with the T50 values increased from 174 °C under dry to 191 °C under wet conditions.

re

For the bimetallic 1:1 Pt-Pd/Al2O3 catalyst, the conversion data are presented in Figure 1. Here

lP

the T50 under wet conditions (152 °C) is slightly lower than under dry conditions (159 °C). Although for both experiments CO oxidation onset occurred at the same temperature (the conversions up to ~5% are identical), the rate of change in conversion after this low conversion

ur na

region varied depending on if water was present.

The catalyst performed better at lower

temperatures in dry conditions, but at higher temperatures (above T50) the inclusion of water led to better performance.

Jo

The turnover frequency (TOF) data for these experiments are shown in Supplemental Information Figure S1. For the dry case, the TOF for the catalysts are very similar, however under wet conditions the TOF trends match the trends/differences observed in conversion. The catalysts, dispersion, loading, and TOF data is presented in Table 1. Qualitatively, the bimetallic catalyst presented aspects from both monometallic catalysts, but in different temperature regions. At low temperature, Pt aspects were noted - low temperature CO 9

oxidation under dry conditions matched those for the monometallic Pt results. When including water in the reaction mixture, the rate of conversion change matched those of Pd.

DRIFTS during CO Oxidation. Example DRIFTS spectra obtained after exposure to CO and oxygen at 66°C are shown in Figure 2. This temperature is below where any oxidation was observed. For the Pd/Al2O3 spectra, Figure 2a, the peak at 2102 cm-1 is assigned to linearly bound

of

CO, and the broad band at 2000-1800 cm-1 to bridged-bound CO19. Based on previous literature,

ro

the higher wavenumber peak in the 2170 cm-1 range is assigned to CO bound to alumina support Lewis acid sites, expected in the 2200-2185 cm-1 range, or the superimposition of CO hydrogen

-p

bonded to surface alumina hydroxyl groups and CO interacting with partially oxidized Pd ions, in

re

the 2170-2130 cm-1 range.8 In this pair of spectra, it is apparent that adding water into the reaction mixture and pretreatment increased linearly bound CO and the species associated with the 2170

lP

cm-1 feature, and led to a decrease in the amount of bridged carbonyls. For the Pt/Al2O3 spectra, Figure 2b, the peak at 2094 cm-1 is assigned to linearly bound CO and

ur na

shoulders at 2060 cm-1 and 2113 cm-1 correspond to co-adsorbed CO and atomic oxygen on a single Pt site.11 The small feature at 1820 cm-1 suggests there was also triply bound CO in the absence of water, and there is no evidence of the doubly bridged CO.19 Both the monometallic catalysts contained less bridged bound CO with water exposure. On the Pt catalyst the CO signals

Jo

without water in the reaction mixture were higher, suggesting more CO adsorbed, while the opposite was observed with the monometallic Pd sample. The Pt-Pd/Al2O3 spectra are shown in Figure 2c. The addition of water to the reaction mixture

led to significantly more CO adsorbed based in signal intensity. No triply bridged carbonyls were observed, but there was a peak at 1965 cm-1 assigned to doubly bridged CO, as was observed with

10

the monometallic Pd catalyst. The linearly adsorbed carbonyl feature for the bimetallic catalyst appears to be composed of two peaks, encompassing 2102-2090 cm-1, here attributed to adsorbed CO on Pt and Pd. One noticeable difference is that when water was absent from the reaction mixture the higher wavenumber peak decreased relative to the lower, indicating that the CO preferentially adsorbed on Pt when water was absent. When water was in the reaction feed, more CO adsorbed onto the Pd, relative to the dry conditions, both consistent with the monometallic

of

trends discussed above. The IR features in the 2200-2130 cm-1 region can be associated with a

ro

few different species, per the literature,8 such as: 2200-2185 cm-1 to CO bound to alumina support Lewis acid sites (Al3+); 2165-2150 cm-1 to CO hydrogen bonded to surface alumina hydroxyl

-p

groups; or CO interacting with fully/partially oxidized Pd ions (2170-2130 cm-1) or Pt ions (2185-

re

2135 cm-1), appearing as peaks moving to lower wavenumber at lower oxidation states. The peak at 2170 cm-1 was larger in magnitude for both the Pd and Pt-Pd catalysts in the presence of water,

lP

while for the Pt, water diminished its size. This feature at 2170 cm-1 may correspond to CO hydrogen bonded to surface alumina hydroxyl groups This aligns with a CO oxidation mechanism

studies.6

ur na

on Pd/Al2O3 where water dissociates and reacts with CO, which is coincident with isotope labeling

CO oxidation mechanisms, with and without water, and the water gas shift (WGS) reaction mechanism, will be discussed briefly in order to summarize possible intermediate species observed

Jo

spectroscopically in the DRIFTS experiments.

First, for the case without water, the

carbonate/carboxylate (CO3/CO2) intermediated CO oxidation mechanism is summarized by the following reactions:10,20,21 𝐶𝑂 + 𝑂2 → 𝐶𝑂3

(1)

𝐶𝑂 + 𝑂 → 𝐶𝑂2

(2)

11

𝐶𝑂3 → 𝐶𝑂2 + 𝑂

(3)

𝐶𝑂2 → 𝐶𝑂2(𝑔)

(4)

Reactants, unless otherwise denoted, represent species adsorbed to the surface. This simplifies the list as it allows us to not assign the formed species as monodentate, bidentate, or as a free ion, as the carbonate can adsorb in these forms, and from the IR spectra we cannot distinguish whether

of

this occurs on the metal sites or on the alumina support. Secondly, the CO oxidation mechanism with water with bicarbonate and carboxyl species

ro

intermediates is shown in reactions 5 through 8,6 and the WGS mechanism is captured by reactions

-p

5 and 9 as follows:22,23

𝐶𝑂 + 𝑂 + 𝑂𝐻 → 𝐶𝑂2 𝑂𝐻

lP

𝐶𝑂𝑂𝐻 + 𝑂𝐻 → 𝐶𝑂2(𝑔) + 𝐻2 𝑂(𝑔)

re

𝐶𝑂 + 𝑂𝐻 → 𝐶𝑂𝑂𝐻

(5) (6) (7) (8)

𝐶𝑂𝑂𝐻 → 𝐶𝑂2(𝑔) + 𝐻

(9)

ur na

𝐶𝑂2 𝑂𝐻 + 𝑂𝐻 → 𝐶𝑂2(𝑔) + 𝐻2 𝑂(𝑔) + 𝑂

The atomic oxygen present in reactions (2) and (6) could be interpreted as either O2 dissociative chemisorption on the catalyst, or comes from the reduction of a metal oxide. Bicarbonate is formed as a reaction intermediate in Eq. 6, and will sit either on the support or metal sites.

A

Jo

formate/carboxyl reaction intermediate is formed in reaction 5, with direct combination of the OH to the CO. Below, COOH is referred to as carboxyl to distinguish it from the HCOO formate species, however, note that the formate species below has also been identified as a WGS intermediate. The formation of the formate species could come about through the following reactions (10-12):

12

𝐶𝑂 + 𝑂𝐻 → 𝐻𝐶𝑂𝑂

(10)

𝐶𝑂 + 𝐻 + 𝑂 → 𝐻𝐶𝑂𝑂

(11)

𝐻𝐶𝑂𝑂 + 𝑂𝐻 → 𝐶𝑂2(𝑔) + 𝐻2 𝑂(𝑔)

(12)

The formate can either form on the support through reaction with alumina OH groups, similar to what occurs on ceria,24,25 or when water is present in the reaction, it can dissociate to form OH

of

on metal sites. The difference between reactions 10 and 5 is that in reaction 10, the OH bond cleaves and the HCOO will be adsorbed in a bidentate configuration with the H sitting

ro

perpendicular to the metal site or support 26.

in the following reactions (reactions 13 and 14):

re

𝐶𝑂 + 𝐻 → 𝐻𝐶𝑂

-p

Finally, any hydrogen present that did not react to form an OH group could itself react with CO

(14)

lP

𝐶𝑂 + 𝐻 → 𝐶𝑂𝐻

(13)

The HCO species represents formyl which was not observed spectroscopically and will not be mentioned further.26 The COH species may indeed have been present, as the IR peak is common

ur na

with many of the other species mentioned.19

In summary, there are seven anticipated/possible intermediate species during CO oxidation as follows: monodentate carbonate, bidentate carbonate, carboxylate, bicarbonate, carboxyl, formate,

Jo

and COH. Spectroscopically, we cannot distinguish whether these species are present on catalytic sites or on the support, but the intermediates themselves can be distinguished, primarily via analysis in the 1800-1100 cm-1 wavenumber range.

We chose an intermediate oxidation

temperature of 162 °C to further characterize intermediates, with an example spectra set from the bimetallic sample shown in Fig 5. It should be noted that the water signal was subtracted from these spectra and so while the 1653 cm-1 feature is very close to water bending at 1637 cm-1 7, we 13

considered it to be not associated with water. The prominent peaks appear between 1258-1228 cm-1 (referred to as 1240 cm-1), 1443-1432 cm-1 (referred to as 1435 cm-1), and 1658-1650 cm-1 (referred to as 1653 cm-1). The exact wavenumbers for the maximum peak heights vary slightly as a function of catalyst and experimental conditions. These peaks appear at varying intensities and ratios across the different samples, which aid in assigning them to surface species and the following logic was used for the assignments.

of

a) A peak at 1435 cm-1 indicates either monodentate or free carbonates,27,28 along with

ro

weaker peaks at 1330 cm-1 for monodentate and 1090-1020 cm-1 for monodentate and free carbonates.

-p

b) Two peaks at 1653 cm-1 and 1240 cm-1 together indicate bidentate carbonates,27,29,30

re

again a weak peak is expected around 1020 cm-1.

c) Carboxylate can be identified by 1435 and 1653 cm-1 appearing together.27,30

lP

d) All three peaks at 1435 cm-1, 1653-1 cm-1 and 1240 cm-1 together indicate bicarbonate species,8 with a weak peak around 1300 cm-1.

ur na

e) A pair of 1240 and 1435 cm-1 peaks correspond to carboxyl.28 f) A peak at 1547 cm-1, which together with a peak at 2960 cm-1 and small shoulder at 1392 cm-1 indicates formate species.8,31 g) Finally the 1240 cm-1 peak occurring alone has been assigned to COH; it has been

Jo

reported on Pt at 1256 cm-1.19

From this, the carbonate peaks in a) and b) correspond to the carbonate formation in reaction 1.

Peaks in the 1100-1000 cm-1 region anticipated for these were not discernible over the noise in the spectra. The peaks mentioned in c) correspond to carboxylate formation in reaction 2. The formation of bicarbonate and carboxyl species mentioned in d) and e) are captured by reactions 5

14

and 6. The formate peaks, from reactions 10 and 11, are given in f). The COH peak given in g) could correspond to CO reacting with H on the surface that has not yet reacted with dissociated oxygen.

Many of these peaks overlap, and so distinguishing between these species

spectroscopically is a challenge, however there are general trends noted, which will be discussed. In order to differentiate between mechanisms and species, three characteristic peak heights were chosen to monitor as a function of temperature; 1240, 1435, and 1635 cm-1, plotted in Figure 4.

of

The relative intensities and changes in these features, based on the analysis provided in a)-g) above,

ro

provide guidance in making the assignments.

The results from the Pd catalyst are shown in Figures 4a, 4d, and 4g. When water was not

-p

present in the reaction, the peak at 1240 cm-1 increased with temperature, while the peaks at 1435

re

and 1653 cm-1 increased as well but to a much lesser extent, and both then plateaued. This suggests that the bicarbonate species (requiring all three peaks) formed from the OH on the support and

lP

then remained on the support, likely at the metal/support interface and did not accumulate further with temperature. This was observed in our previous study and the formation of carbonate was

ur na

found to be an inhibiting surface intermediate.11 The same occurs with the carboxyl formed (1240 cm-1 and 1435 cm-1). The more significant increase in the 1240 cm-1 feature alone demonstrates COH species formation. This shows that OH species on the support and adjacent to the metal particles, even without water in the gas phase, react with CO to form the COH species, which in

Jo

turn remain inactive in the absence of larger amounts of OH that would form from water dissociation when water was added to the reaction mixture. The build-up of these species may reduce the surface area available for CO oxidation with oxygen and explain why the performance for the monometallic Pd catalyst in the absence of water in the reaction mixture was poorer. When water was present, the 1653 cm-1 peak did not change much compared to that without water, and

15

there was a marked decrease in the 1435 and 1240 cm-1 peaks associated with carboxyl; so with water in the reaction mixture, the carboxyl was able to react with OH groups from dissociated water. When water is introduced to the reaction mixture, the oxidation mechanism for CO follows carboxyl and bicarbonate intermediated mechanisms, whereas for the dry condition the CO oxidation mechanism follows mainly the carbonate/carboxylate mechanism. A similar CO oxidation mechanism over Au/TiO2 and Au/Al2O3 with water was reported where water binding

of

on the support near the Au particles was found to be part of the mechanism, with CO being

ro

oxidized through a carboxyl intermediated mechanism.32 This study also found carbonate formation to be inhibiting and more pronounced over the Al2O3 support than the TiO2 support.

-p

The peaks at 1240 cm-1, 1435 cm-1, and 1653 cm-1 are plotted as a function of inlet gas

re

temperature for the Pt catalyst in Figures 4b, 4e and 4h, respectively. Again, since all three peaks are present in the absence of water, bicarbonate species formed, as well as formate species based

lP

on the presence of a feature at 1547 cm-1 (f in the list above, spectra shown in Supplementary Information S2), indicating a bicarbonate mechanism requiring OH groups. When water was

ur na

present in the reaction mixture, none of these peaks increased, or appeared, as a function of temperature. When water was not present, there was an intermediate temperature range where the 1435 cm-1 and 1653 cm-1 peaks increased, much like with the Pd catalyst, except that these peaks reached a maximum and then decreased as a function of temperature. At 166°C the 1435 cm-1 was

Jo

3.4x greater with the Pt catalyst compared to the Pd catalyst, while the 1653 cm-1 peak was only 2x greater. The 1240 cm-1 peak height was also greater at that temperature for the Pt catalyst than the Pd catalyst, by 2.4x. This demonstrates the higher 1435 cm-1 for the Pt catalyst is due to monodentate carbonate rather than carboxyl species, in addition to the bicarbonate or bidentate carbonate species in this intermediate temperature range. At higher temperatures, the 1240 cm-1

16

peak remained while the other two peaks decreased, again suggesting COH accumulation. From these observations, CO oxidation on Pt appears to involve more carbonate intermediates than over the Pd catalyst. Note that on the Pt catalyst, unlike the Pd catalyst, with water in the reaction no bicarbonate was observed. Finally, on the Pt-Pd catalyst, the peak heights at 1240 cm-1, 1435 cm-1 and 1653 cm-1 are plotted as a function of inlet gas temperature for the Pt-Pd catalyst in Figure 4c, 4f and 4i, respectively.

of

Figure 3 shows large 1435 cm-1 and 1653 cm-1 peaks at the beginning, with a small peak at 1240

ro

cm-1, indicating carboxylate, monodentate carbonate, and bicarbonate. In this case, only small 1240 cm-1 features were observed, with no observed difference with or without water and no

-p

significant increase as a function of temperature. This suggests that bicarbonates, carboxyls, and

re

COH species did not accumulate on the surface of the bimetallic sample with increasing temperature. When water was not present, the 1435 and 1653 cm-1 peaks appeared at intermediate

lP

temperatures, much like what was observed for the Pt and Pd catalysts, except with no accompanying 1240 cm-1 peak. In terms of relative peak heights, picking 166 °C again, without

ur na

water in the pretreatment or reaction, the value for the 1435 cm-1 peak was similar to that of the Pt catalyst and 3.8x that of the Pd catalyst. The 1635 cm-1 peak was 3.6x greater than what was observed on the Pd catalyst. This is a similar increase to that noted for the Pt catalyst, and suggests that these two peaks are related to the same species, which further strengthens the assignment to

Jo

carboxylate.

To summarize these observations, it appears that that bicarbonate, bidentate carbonate, and

carboxyl species formed on the monometallic Pd surface. On the Pt catalyst, under dry conditions, some bicarbonate, monodentate carbonate, and formate species formed. Both of the monometallic catalysts showed evidence of COH formation. For the Pt-Pd catalyst, more carboxylate and

17

monodentate carbonates formed, and only a small amount of the bicarbonate. Also, the lower 1240 cm-1 peak on the Pt-Pd catalyst compared to the monometallic samples suggests the bimetallic species suppressed COH accumulation. This could explain why the Pt-Pd catalyst resulted in the greatest rate of conversion change as a function of temperature during the bench scale reactor experiments. At the same time, the accumulation of the carboxyl and bicarbonate surface species, as well as the COH, in the absence of water would explain the inhibition observed.

of

For the Pt-Pd catalyst, the water did not have as much of a beneficial effect on CO conversion,

ro

and the DRIFTS show the bridged CO species was not present. The addition of Pt to form the bimetallic inhibited CO from binding in a bridged mode, supported by the suppression of bidentate

-p

carbonate formation during CO oxidation on the Pt-Pd catalyst. However, for the Pt-Pd catalyst,

re

the CO was still able to adsorb either on dissociated OH on the metal or OH from the support/particle interface when water was in the reaction feed, unlike on the Pt catalyst, and so the

lP

greater rate of change of conversion associated with the bicarbonate/carboxyl mechanism was still attained when water was present in the reaction mixture. Also, when water was not present in the

ur na

reaction mixture, the OH region 3800-3000 cm-1 decreased as a function of temperature for all samples. The role of the water appears to be to increase CO adsorption at the interface between the support and Pd and Pt-Pd particles, replenish OH groups on the support, and also dissociate on the metal sites themselves. For the Pt, the adsorption of water on the metal sites when water was

Jo

present in the reaction mixture causes inhibition. For the Pd, the higher amount of CO at the interface seems to be relevant in the CO oxidation mechanism. These two factors lead to the intermediate behavior seen on the bimetallic catalyst.

18

C3H6 Oxidation. Propylene oxidation reactions were also characterized. In a previous study,11 the partial oxidation products observed during propylene oxidation over different Pt:Pd catalysts were as follows: CO, formaldehyde, ethylene, acetaldehyde, acetic acid, and acetone. In such high oxygen concentrations, the complete oxidation of propylene to CO2 and H2O shown below in reaction 15 would be expected. 9 2

𝑂2(𝑔) → 3𝐶𝑂2(𝑔) + 3𝐻2 𝑂(𝑔)

(15)

of

𝐶3 𝐻6(𝑔) +

As mentioned, partial oxidation products do form even in the abundance of oxygen, due to high Previous IR experiments have shown that propylene

ro

surface concentrations of propylene.

-p

adsorbed as propylidyne, as shown in reactions 16a, 16b and 16c as follows: 𝐶3 𝐻6(𝑔) + 𝑂 → ≡ 𝐶 − 𝐶𝐻2 − 𝐶𝐻3 + 𝑂𝐻

(16a) (16b)

2𝐶3 𝐻6(𝑔) → 2( ≡ 𝐶 − 𝐶𝐻2 − 𝐶𝐻3 ) + 𝐻2

(16c)

lP

re

2𝐶3 𝐻6(𝑔) + 𝑂2 → 2( ≡ 𝐶 − 𝐶𝐻2 − 𝐶𝐻3 ) + 𝐻2 𝑂 + 𝑂

This configuration, unlike π or σ -bound propylene, allows the propylene to sit perpendicular to

ur na

the surface and therefore leads to the highest surface coverage, but requires a hydrogen to be abstracted, which may react with molecular oxygen to form water or with atomic oxygen to form OH, or alternatively at low oxygen coverages may lead to H2 formation. At higher temperatures, this species likely rearranges to a 1-methylvinyl species (where the

Jo

central carbon is now doubly bound to the surface) as shown in reaction 17. ≡ 𝐶 − 𝐶𝐻2 − 𝐶𝐻3 → −𝐶𝐻2 − 𝐶 − 𝐶𝐻3

(17)

At low oxygen surface concentrations further dehydrogenation of the alyllic carbon occurs to form water, reactions 18a and 18b. −𝐶𝐻2 − 𝐶 − 𝐶𝐻3 + 𝑂 → −𝐶𝐻 − 𝐶 − 𝐶𝐻3 + 𝑂𝐻

19

(18a)

−𝐶𝐻 − 𝐶 − 𝐶𝐻3 + 𝑂𝐻 → −𝐶 − 𝐶 − 𝐶𝐻3 + 𝐻2 𝑂

(18a)

However, the production of acetone is possible via oxygen attack of the vinyl carbon, reaction 19. Note that the formation of acetone requires regaining the H that was lost during adsorption and so the reaction has been written with OH, indicating dissociative adsorption of the produced water. −𝐶𝐻2 − 𝐶 − 𝐶𝐻3 + 𝑂𝐻 → (𝐶𝐻3 )2 𝐶 = 𝑂

(19)

of

Ethylene formation comes from cleavage of a carbon-carbon bond, the allylic carbon (the carbon attached to the metal site), to form ethylidyne and CO via another oxygen attack, reaction 20. This

ro

ethylidyne can rearrange and react with H, then desorb to form ethylene.

(20)

-p

−𝐶 − 𝐶 − 𝐶𝐻3 + 𝑂 → 𝐶𝑂 + ≡ 𝐶 − 𝐶𝐻3

In the previous IR study, surface acetate was observed as was product acetic acid, which would

(22) and acetic acid (23) as follows:

lP

≡ 𝐶 − 𝐶𝐻3 + 𝑂 → = 𝐶𝑂 − 𝐶𝐻3

re

involve the reaction of ethylidyne with oxygen as in reaction 21 and subsequent reactions to acetate

(21) (22)

= 𝐶𝑂 − 𝐶𝐻3 + 𝑂𝐻 → 𝐶𝐻3 − 𝐶𝑂𝑂𝐻

(23)

ur na

= 𝐶𝑂 − 𝐶𝐻3 + 𝑂 → 𝐶𝐻3 − 𝐶𝑂2

Acetaldehyde was observed as a product in the gas phase, but not on the surface. The acetaldehyde may form as shown in the following reaction. ≡ 𝐶 − 𝐶𝐻3 + 𝑂𝐻 → 𝐶𝐻3 − 𝐶𝐻𝑂

Jo

(24)

Also in the previous study, acrolein was not observed, nor was it observed in this study. Here,

product formaldehyde was observed, likely through reactions 25-27. The discussed reaction mechanism was deduced from DRIFTS results that were obtained in the absence of water, and since only small changes in the observed outlet concentrations involving the same intermediates were observed in this study, DRIFTS experiments were not again conducted. There was also 20

formation of methane, both with and without water in the reaction, observed over the Pd and PtPd catalysts (shown in supplemental information). = 𝐶𝑂 − 𝐶𝐻3 → −𝐶𝐻3 + 𝐶𝑂

(25)

−𝐶𝐻3 + 𝑂 → 𝐶𝐻2 + 𝑂𝐻

(26)

𝐶𝐻2 + 𝑂 → 𝐶𝐻2 𝑂

(27)

of

Having gone through the previously reported propylene oxidation reaction mechanism over PtPd catalysts in the absence of water, the effect of adding water is now evaluated. Note that in the

ro

above reaction scheme, OH participated in reactions 19, 23 and 24, so one might expect higher

-p

concentrations of acetone, acetic acid, and acetaldehyde by adding water due to higher surface OH concentrations. There is also OH or water formation in reactions 16 and 18, which may imply that

re

water could impact the adsorption and dehydrogenation of propylene.

With the inclusion of water in the reaction mixture, steam reforming (SR) reactions could also

lP

become important, however, SR typically occurs at temperatures above 400°C for Pt/Al2O3 33, and thus were not considered, as they are typically neglected during low temperature oxidation 5,34.

ur na

As shown in Figure 5a, C3H6 oxidation over the Pd/Al2O3 sample was impacted by water addition (dry vs. wet). The T50 increased from 190 °C to 214 °C from dry to wet. Partial oxidation product formation; CO, ethylene, and acetone are shown in Figures 5b, 5c, and 5d, respectively. The most abundant partial oxidation intermediate was acetone, and it and ethylene appeared in the

Jo

highest concentrations for the dry experiment. On the other hand, the most CO formed in the wet experiment. This is counterintuitive as the Pd catalyst was the better CO oxidation catalyst with water in the reaction mixture as described above, so it is curious that this partial oxidation product was not more easily oxidized in the presence of water. The peak partial oxidation product formation coincided with propylene conversion in the 45-60% range. The peak CO formation 21

occurred prior to that of ethylene formation.

Other partial oxidation products such as

formaldehyde, acetaldehyde, acetic acid, and acrolein, usually anticipated from propylene oxidation, were also measured. As mentioned, acrolein was not observed and formaldehyde was formed at concentrations below 2 ppm over all catalysts and conditions and its concentrations are therefore not shown. Acetaldehyde and acetic acid formation were on the order of 5 ppm, and results are provided in supplemental information for all catalysts in Figures S3-5. Acetic acid

of

formation only occurred with water in the reaction mixture, but not on the Pt catalyst, and its

ro

concentrations are also shown in the supplemental information. There were no interesting trends with the addition of water on acetaldehyde formation.

-p

Relating these results back to the mechanism discussed, there was an increase in acetone,

re

ethylene, CO and acetic acid intermediate concentrations when water was introduced to the reaction mixture, and no effect on acetaldehyde formation. The increase in ethylene formation

lP

was much less than the increase in CO formation, which indicates that this CO may not originate from the first C-C bond cleavage in reaction 20, but from the further partial oxidation of C2 or C1

ur na

partial oxidation products. Acetone formation, reaction 19, was greater with water added to the reaction mixture, and as this reaction involves OH, increased water dissociation would lead to more of this product. There was also more acetic acid production when water was in the reaction mixture (reaction 23), also consistent with higher OH surface concentrations. The increase in OH

Jo

surface concentrations and partial oxidation product formation likely causes the increase in light off temperatures observed in the presence of water, due to surface intermediate species buildup. In contrast to the Pd catalyst, the presence of water in the reaction actually reduced the

temperature for C3H6 light off over Pt/Al2O3, with a T50 decrease from 273 to 263 °C, as shown in Figure 6a. In terms of the partial oxidation intermediates, ethylene was not observed during any

22

experiments and acetone was only observed when water was present in the reaction, but only at very low concentrations, < 2 ppm. In previous work11, it was hypothesized that on the Pt surface, the indirect propylene oxidation mechanism is favored (reactions 18 and 20), where hydrogen is abstracted from the surface to produce ethylene and CO instead of acetone (reaction 19). Here, water impacted the balance between direct and indirect oxidation occurring on the surface. With the addition of water,

of

dissociation occurs, and the OH can then react with adsorbed propylene to produce acetone,

ro

however still at very low concentrations. Note that the previous study showed acetone on the surface but not in the gas phase, here with water added there were higher surface acetone

-p

concentrations formed such that some desorbed from the catalyst surface. CO formation, shown

re

in Fig 10b, was lower in the presence of water, which is further consistent with the shift towards acetone production away from the dehydrogenation reaction (18) leading to CO formation (20).

lP

The higher prevalence of the mechanism towards acetone in this case led to a lower light off temperature, even though on the Pd catalyst this path inhibited performance. This alludes to a

ur na

desired balance between these mechanisms. Also, the acetaldehyde concentration decreased when water was in the reaction mixture, even though OH is a reactant in reaction 24. This suggests that on the Pt catalyst, the OH abstracts hydrogen from ethylidyne, or reaction 22 to acetate, or complete oxidation is favored. In our previous IR study, there was more ethylidyne on the surface

Jo

of the Pt catalyst and less acetate, so the reaction of OH with ethylidyne to acetaldehyde may have a higher activation barrier on Pt. For the bimetallic Pt-Pd/Al2O3 catalyst, the presence of water in the reaction mixture inhibited

propylene oxidation light off as was observed for the Pd catalyst, as shown in Figure 7a. The addition of water increased T50 only slightly, from 189 °C to 194 °C. The inhibition by water

23

follows the Pd catalyst trend, but the inhibition extent is significantly decreased, due to the influence of Pt. CO, ethylene, and acetone oxidation by-product concentrations are plotted in Figures 7b, 7c and 7d. Overall, the bimetallic catalyst resulted in similar levels of partial oxidation products as the monometallic Pd catalyst, except for CO, along with very little C3H6 oxidation inhibition by water. CO formation was lower than that observed from either Pt or Pd, and there was more CO formed when water was present in the reaction mixture.

The higher CO

of

concentration with the addition of water was also observed on the Pd catalyst. And similar

ro

ethylene and acetone formation levels as were obtained with the Pd catalyst were observed. Interestingly even though the T50 shifted upward when water was in the reaction mixture, the

-p

conversions in both cases up to 20% were very similar. This is similar to what was observed on

re

the Pt catalyst where water had a small effect on the lower conversion. The greater change of C3H6 conversion with temperature over the Pt-Pd catalyst relative to either the Pt or Pd catalysts may be

propylene oxidation.

lP

related to the CO formation, as there was less CO formed on the bimetallic catalyst to inhibit

ur na

Relating this back to the mechanism, for the Pt-Pd catalyst the higher outlet concentrations of CO in the presence of water, like on the Pd sample, can be related to the partial oxidation of C2 and C1 products, instead of being produced via reaction 20, which led to ethylene and CO formation. However, there was still less CO formation than that from the monometallic Pd

Jo

catalyst. When water was in the reaction mixture, there was less acetaldehyde produced and more acetic acid, indicating the increased OH on the surface influenced the ratio of reactions 23 and 24. This is counterintuitive, as one would expect that an increase in surface OH concentration would lead to more acetaldehyde formation in reaction 24 since 23 requires the reaction of the same species with atomic oxygen before reaction 23 can take place. However, for the Pt catalyst with

24

water in the reaction mixture, there was also lower acetaldehyde production. Chemistry associated with both metals is therefore observed; with Pd, more acetic acid production with water in the reaction, and with Pt, lower acetaldehyde production with water in the reaction. As a final note, under real exhaust conditions, the oxidation catalyst would be exposed to both CO and hydrocarbons at the same time. Many studies have shown that over Pt-Pd catalysts, it is important to oxidize the CO before hydrocarbon oxidation can take place. Therefore, we would

of

expect that in the case of co-oxidation of CO and propylene with the addition of water, that while

ro

there would be inhibition due to the competitive adsorption of CO and propylene on the catalyst surface, it would be less pronounced over the Pd and 1:1 Pt:Pd catalysts due to the promotion of

-p

CO oxidation with water. However, at the same time we would expect that the water would further

re

inhibit oxidation of the hydrocarbons on the Pd catalyst. On the other hand, on the Pt where CO oxidation was inhibited by the addition of water to the reaction mixture, we would expect that the

lP

benefit to the propylene lightoff with the addition of water over the Pt catalyst would not be

ur na

observed.

CONCLUSIONS

The effect of water on CO and propylene oxidation under simulated RCCI exhaust conditions was investigated. Water lowered the CO light-off temperature on the monometallic Pd catalyst

Jo

while it inhibited CO oxidation with the monometallic Pt catalyst. With the Pt-Pd catalyst, water accelerated the reaction at high temperatures, but had no impact on low temperature/low conversion CO oxidation. DRIFTS characterization results showed that there was more CO adsorption on the Pd and Pt-Pd catalysts when water was present. Surprisingly the opposite effect was observed on the Pt catalyst, the water appeared to compete with CO adsorption which may account for the inhibition observed. 25

For propylene oxidation, the inclusion of water into the reaction mixture had trivial effects on the partial oxidation product distributions; slightly more CO was produced over the Pd and Pt-Pd catalysts when water was in the reaction mixture and slightly less was produced over the Pt catalyst. The water did have an impact on the light-off performance. On the Pd catalyst, when water was in the reaction, propylene oxidation light-off was inhibited. For Pt, including water lowered the propylene light-off temperature. Water inhibited CO oxidation, so this may suggest

ro

of water in the reaction and pretreatment mixtures was observed.

of

water acted directly as a propylene oxidant. For the Pt-Pd catalyst, little inhibition via inclusion

-p

Declaration of interests

re

The authors declare that they have no known competing financial interests or personal

Author Contribution Statement

lP

relationships that could have appeared to influence the work reported in this paper.

ur na

Melanie Hazlett: Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Visualization; Writing original draft

Jo

William Epling: Formal analysis; Funding acquisition; Project administration; Supervision; Writing - review & editing

ACKNOWLEDGEMENTS We thank the US Department of Energy and the National Science Foundation (CBET 1258688) for financial support.

26

REFERENCES 1

A. T. Gremminger, H. W. Pereira de Carvalho, R. Popescu, J.-D. Grunwaldt and O. Deutschmann, Catal. Today, 2015, 1–11. P. Gélin, L. Urfels, M. Primet and E. Tena, Catal. Today, 2003, 83, 45–57.

3

D. L. Mowery, M. S. Graboski, T. R. Ohno and R. L. McCormick, Appl. Catal. B Environ.,

of

2

1999, 21, 157–169.

H. J. Kwon, J. H. Baik, Y. T. Kwon, I.-S. Nam and S. H. Oh, Chem. Eng. J., 2008, 141,

ro

4

5

-p

194–203.

O. Mihai, A. Fathali, X. Auvray and L. Olsson, Appl. Catal. B Environ., 2014, 160–161,

6

re

480–491.

R. Caporali, S. Chansai, R. Burch, J. J. Delgado, A. Goguet, C. Hardacre, L. Mantarosie

7

F. Liang, H. Zhu, Z. Qin, H. Wang, G. Wang and J. Wang, Catal. Letters, 2008, 126, 353–

ur na

360. 8

lP

and D. Thompsett, Appl. Catal. B Environ., 2014, 147, 764–769.

G. Busca, E. Finocchio and V. S. Escribano, Appl. Catal. B Environ., 2012, 113–114, 172– 179.

A. B. Mhadeshwar and D. G. Vlachos, J. Phys. Chem. B, 2004, 108, 15246–15258.

10

M. Moses-Debusk, M. Yoon, L. F. Allard, D. R. Mullins, Z. Wu, X. Yang, G. Veith, G. M.

Jo

9

Stocks and C. K. Narula, J. Am. Chem. Soc., 2013, 135, 12634–12645.

11

M. J. Hazlett, M. Moses-Debusk, J. E. Parks, L. F. Allard and W. S. Epling, Appl. Catal. B Environ., 2017, 202, 404–417.

12

V. Y. Prikhodko, S. J. Curran, T. L. Barone, S. A. Lewis, J. M. Storey, K. Cho, R. M. Wagner and J. E. Parks, in American Society of Mechanical Engineers International 27

Mechanical Engineering Conference and Exposition, 2011, pp. 273–278. 13

V. Y. Prikhodko, S. J. Curran, T. L. Barone, S. A. Lewis, J. M. Storey, K. Cho, R. M. Wagner and J. E. Parks, SAE Int. J. Fuels Lubr., 2012, 3, 946–955.

14

M. Moses-DeBusk, S. J. Curran, S. A. Lewis, R. M. Connatser and J. M. E. Storey, Emiss. Control Sci. Technol., 2019, 5, 225–237.

15

J. Farrell, J. Holladay and R. Wagner, "Fuel Blendstocks with the Potential to Optimize

of

Future Gasoline Engine Performance: Identification of Five Chemical Families for Detailed Evaluation," Technical Report, U.S. Department of Energy, Washington, DC.,

ro

2018. DOE/GO-102108-4970.

M. Skoglundh, L. O. Lӧwendahl and J. Ottersted, Appl. Catal., 1991, 77, 9–20.

17

The Advanced Combustion and Emission Control (ACEC) Technical Team Low-

-p

16

Temperature Aftertreatment Group, Aftertreatment Protocols for Catalyst Characterization

S. B. Kang, M. Hazlett, V. Balakotaiah, C. Kalamaras and W. Epling, Appl. Catal. B

lP

18

re

and Performance Evaluation: Low-Temperature Oxidation Catalyst Test Protocol, 2015.

Environ., , DOI:10.1016/j.apcatb.2017.05.057. T. Iwasita and F. C. Nart, Electroanal. Chem., 1991, 317, 291–298.

20

A. D. Allian, K. Takanabe, K. L. Fujdala, X. Hao, T. J. Truex, J. Cai, C. Buda, M. Neurock

ur na

19

and E. Iglesia, J. Am. Chem. Soc., 2011, 133, 4498–4517. G. Djéga-Mariadassou and M. Boudart, J. Catal., 2003, 216, 89–97.

22

C. M. Kalamaras, G. G. Olympiou and A. M. Efstathiou, Catal. Today, 2008, 138, 228–

Jo

21

234.

23

G. G. Olympiou, C. M. Kalamaras, C. D. Zeinalipour-Yazdi and A. M. Efstathiou, Catal. Today, 2007, 127, 304–318.

24

G. Jacobs, U. M. Graham, E. Chenu, P. M. Patterson, A. Dozier and B. H. Davis, J. Catal., 2005, 229, 499–512. 28

25

R. Burch, Phys. Chem. Chem. Phys., 2006, 8, 5483–5500.

26

O. Pozdnyakova, D. Teschner, A. Wootsch, J. Kröhnert, B. Steinhauer, H. Sauer, L. Toth, F. C. Jentoft, A. Knop-Gericke, Z. Paál and R. Schlögl, J. Catal., 2006, 237, 17–28.

27

A. A. Davydov, Infrared Spectroscopy of Adsorbed Species on the Surfaces of Transition Metal Oxides, John Wiley & Sons, Chichester, 1990. Y. Zhu, H. Uchida and M. Watanabe, Langmuir, 1999, 15, 8757–8764.

29

I. M. Hamadeh and P. R. Griffiths, Appl. Spectrosc., 1987, 41, 682–688.

30

K. I. Hadjiivanov and G. N. Vayssilov, Adv. Catal., 2002, 47, 307–511.

31

M. M. Schubert, H. A. Gasteiger and R. Jürgen Behm, J. Catal., 1997, 172, 256–258.

32

J. Saavedra, C. J. Pursell and B. D. Chandler, J. Am. Chem. Soc., 2018, 140, 3712–3723.

33

C.-H. Lee and Y.-W. Chen, Ind. Eng. Chem. Res., 1997, 36, 1498–1506.

34

J. R. Rostrup-Nielsen, in Catalytic Steam Reforming, eds. J. R. Anderson and M. Boudart,

re

-p

ro

of

28

Jo

ur na

lP

Springer-Verlag, Berlin, 1984, pp. 1–118.

29

TABLES Catalyst Pt:Pd molar PGM ratio

Dispersion CO oxidation TOF CO oxidation TOF

Loading (%)

(s-1) at 110 °C dry (s-1) at 110 °C wet

(wt%)

condition

condition 13.7

0:1

0.55

26.2

11.3

1:0

1.00

5.8

20.3

1:1

0.78

21.7

10.7

Jo

ur na

lP

re

-p

Pt-Pd

4.4

ro

Pt

of

Pd

30

6.0

FIGURES

1:1 Pt:Pd/Al2O3 CO Oxidation H2O Effect

80

Dry Pd Wet Pd Dry Pt Wet Pt Dry Pt-Pd

60

of

40 20

ro

CO Conversion [%]

100

0 100

120

140 160 180 Inlet Gas Temperature [°C]

200

220

-p

80

re

Figure 1. CO conversion over the Pd/Al2O3, Pt/Al2O3, and Pt-Pd/Al2O3 catalysts during the

lP

temperature programmed oxidation (TPO) experiment as a function of inlet gas temperature. Reaction conditions 3000 ppm CO, 10% O2, 0 or 5% H2O, in balance N2 and pretreatment

Jo

ur na

conditions 14% O2, 0 or 5% H2O, in balance N2.

31

Dry Wet 1970 cm-1 bridged CO

0.004 0.002

2100 2000 Wavenumber [cm-1]

2090 cm-1 linear CO

b) Pt

1900

Dry Wet

0.012 0.008

-p

2170 cm-1

of

2200

0 1800

ro

2300

KM Units

2170 cm-1

KM Units

a) Pd

0.006

2090 cm-1 linear CO

2170 cm-1

2200

Jo

2300

2090 cm-1 linear CO

ur na

c) Pt-Pd

2100 2000 Wavenumber [cm-1]

1900

Dry Wet

1970 cm-1 bridged CO

2100 2000 Wavenumber [cm-1]

0.008 0.006 0.004 0.002

KM Units

2200

0 1800

lP

2300

re

0.004

0 1800

1900

Figure 2. DRIFTS spectra obtained at 66°C with 98 scans at 4 cm-1 resolution in the 2300-1800 cm-1 region for a) Pd, b) Pt, c) Pt-Pd catalysts; reaction conditions 3000 ppm CO, 10% O2, 0 or 5% H2O, in balance N2 and pretreatment conditions 14% O2, 0 or 5% H2O, in balance N2.

32

0.14 Dry Wet

1435 cm-1

0.06 1240 cm-1

KM Units

0.1

1653 cm-1

1700

1600

1500

1400

1200 -0.02

-p

Wavenumber [cm-1]

1300

ro

1800

of

0.02

Figure 3. DRIFTS spectra obtained at 162°C with 98 scans at 4 cm-1 resolution in 1800-1200 cmregion for the Pt-Pd catalyst; reaction conditions 3000 ppm CO, 10% O2, 0 or 5% H2O, in balance

re

1

0

80 180 280 Temperature [°C] RX DryO2 PT O2 RX O2 H2O PT O2 H2O Wet

c) 1240 cm-1 Pt-Pd 0.3

Peak Height

Peak Height

0.1

b) 1240 cm-1 Pt 0.3

ur na

0.2

Jo

Peak Height

a) 1240 cm-1 Pd 0.3

lP

N2 and pretreatment conditions 14% O2, 0 or 5% H2O, in balance N2.

0.2 0.1 0

0.2 0.1 0

80

180 280 Temperature [°C] Dry Wet

33

80

180 280 Temperature [°C] Dry Wet

f) 1435 cm-1 Pt-Pd 0.09

Peak Height

e) 1435 cm-1 Pt 0.09

Peak Height

0.04

0.04

280

180

280

Temperature [°C] Dry Wet

ur na

Temperature [°C] Dry Wet

-0.02 80

re

180

0.06

lP

-0.02 80

of

Peak Height

0.06

i) 1653 cm-1 Pt-Pd 0.14

ro

h) 1653 cm-1 Pt 0.14

-p

g) 1653 cm-1 Pd 0.14

0.04

-0.01 80 180 280 Temperature [°C] Dry Wet

-0.01 80 180 280 Temperature [°C] Dry Wet

Peak Height

-0.01 80 180 280 Temperature [°C] Dry Wet

Peak Height

Peak Height

d) 1435 cm-1 Pd 0.09

0.06

-0.02 80

180

280

Temperature [°C] Dry Wet

Figure 4. DRIFTS spectra for 1240 cm-1 for a) Pd, b) Pt, and c) Pt-Pd; 1435 cm-1 or d) Pd, e) Pt, and f) Pt-Pd; and 1653 cm-1 for g) Pd, h) Pt, and i) Pt-Pd/Al2O3 catalysts; reaction conditions 3000 ppm CO, 10% O2, 0 or 5% H2O, in balance He and N2 and pretreatment conditions 14% O2, 0 or

Jo

5% H2O, in balance He.

34

a)

100

60 Dry Wet

40

of

C3H6 Conversion [%]

80

ro

20 0 150

200 Inlet Gas Temperature [°C]

250

300

re

-p

100

lP

30

20

10

Jo

0

Dry Wet

ur na

CO Concentration [ppm]

b) 40

100

150

200 Inlet Gas Temperature [°C]

35

250

300

30

20 Dry Wet

10

of

Ethylene Concentration [ppm]

c) 40

150

200 Inlet Gas Temperature [°C]

re lP

80 60 40

0

ur na

Acetone Concentration [ppm]

(d) 100

20

250

300

-p

100

ro

0

Jo

100

150

200 Inlet Gas Temperature [°C]

Dry Wet

250

300

Figure 5. Byproduct concentrations measured when testing the Pd/Al2O3 catalyst; reaction conditions 1500 ppm C3H6, 10% O2, 0 or 5% H2O, in balance N2 and pretreatment conditions 14% O2, 0 or 5% H2O, in balance N2. a) propylene conversion, b) CO concentration, c) ethylene concentration, d) acetone concentration. 36

a)

100

Dry Wet

C3H6 Conversion [%]

80 60 40

of

20

180

230 Inlet Gas Temperature [°C]

b) 40

lP

20

10

0

330

re

Wet

ur na

CO Concentration [ppm]

Dry 30

280

-p

130

ro

0

Jo

130

180

230 Inlet Gas Temperature [°C]

280

330

Figure 6. Byproduct concentrations measured when testing the Pt/Al2O3 catalyst; reaction conditions 1500 ppm C3H6, 10% O2, 0 or 5% H2O, in balance N2 and pretreatment conditions 14% O2, 0 or 5% H2O, in balance N2. a) propylene conversion, b) CO concentration.

37

a)

100

C3H6 Conversion [%]

80 Dry

60

Wet 40

of

20

150

200 Inlet Gas Temperature [°C]

20

0

300

re lP

30

ur na

CO Concentration [ppm]

b) 40

10

250

-p

100

ro

0

Jo

100

150

200 Inlet Gas Temperature [°C]

38

Dry Wet

250

300

30

20

Dry Wet

10

of

Ethylene Concentration [ppm]

c) 40

150

200 Inlet Gas Temperature [°C]

60 40

0

300

re lP

80

ur na

Acetone Concentration [ppm]

d) 100

20

250

-p

100

ro

0

Jo

100

150

200 Inlet Gas Temperature [°C]

Dry Wet

250

300

Figure 7. Byproduct concentrations measured when testing the 1:1 Pt:Pd/Al2O3 catalyst; reaction conditions 1500 ppm C3H6, 10% O2, 0 or 5% H2O, in balance N2 and pretreatment conditions 14% O2, 0 or 5% H2O, in balance N2. a) propylene conversion, b) CO concentration, c) ethylene concentration, d) acetone concentration. 39

40

of

ro

-p

re

lP

ur na

Jo