Cu Model Catalyst Dynamics and CO Oxidation Kinetics Studied by

Subscriber access provided by UNIV OSNABRUECK

Article

Cu model catalyst dynamics and CO-Oxidation kinetics studied by simultaneous in situ UV-Vis and mass spectroscopy Yibin Bu, J. W. (Hans) Niemantsverdriet, and Hans O. A. Fredriksson ACS Catal., Just Accepted Manuscript • DOI: 10.1021/acscatal.5b02861 • Publication Date (Web): 14 Mar 2016 Downloaded from http://pubs.acs.org on March 18, 2016

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

ACS Catalysis is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

1

ACS Catalysis

Cu model catalyst dynamics and CO-

2

Oxidation kinetics studied by simultaneous in situ

3

UV-Vis and mass spectroscopy

4

Yibin Bu, a J.W. (Hans) Niemantsverdriet, b,c Hans O.A. Fredriksson a ,b*

5

a. Laboratory for Physical Chemistry of Surfaces, Eindhoven University of Technology, P.O.

6

Box 513, 5600 MB Eindhoven, The Netherlands

7

b. SynCat@DIFFER, Syngaschem BV, P.O. Box 6336, 5600 HH Eindhoven, The

8

Netherlands

9

c. SynCat@Beijing, Synfuels China Technology Co., Ltd., Huairou, P.R. China

10

ABSTRACT: The oxidation state of Cu nanoparticles during CO oxidation in CO+O2 gas

11

mixtures was sensitively monitored via the localized surface plasmon resonances. A micro

12

reactor, equipped with in situ UV-Vis and mass spectrometry, was developed and used for the

13

measurements. The Cu nanoparticles of ~30 nm average diameter were supported on optically

14

transparent, planar quartz wafers. The aim of the study is two-fold, i) to demonstrate the

15

performance and usefulness of the set-up and ii) to use the combined strength of model catalysts

16

and in situ measurements to investigate the correlation between the catalyst oxidation state and

17

its reactivity. Metallic Cu is significantly more active than both Cu (I) and Cu (II) oxides. The

18

metallic Cu-phase is only maintained under conditions where close to full oxygen conversion 1

ACS Paragon Plus Environment

ACS Catalysis

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

1

is achieved. This implies that kinetic measurements, aimed at determining the apparent

2

activation energy for metallic Cu under realistic steady state conditions, are difficult or

3

impossible to perform.

4

5

6

KEYWORDS: Cu model catalyst, in situ UV-Vis and mass spectroscopy, oxidation state of Cu

7

catalyst, CO oxidation, kinetic measurements

8

1. INTRODUCTION

9

Properties of Cu-based catalysts are of particular interest in that they are applied in many

10

different reactions, e.g., methanol synthesis, water-gas shift and methane reforming. There is

11

still much interest and controversy regarding the oxidation state of Cu catalysts, especially

12

under reaction conditions. For example, the correlation between reactivity and oxidation states

13

is not well established because of the complexity of the catalysts and reaction conditions, and

14

because of the difficulty to monitor the oxidation states of the catalysts, while in use.

15

Carbon monoxide oxidation is frequently used as a test reaction to study basic properties of

16

catalysts 1–19 and abundant work concerning CO oxidation on Cu catalysts has been performed

17

1,6–19

18

various orientations in UHV conditions 6,17–19. It is proposed that high oxygen coverages retard

19

the reaction rate and that the surface reaction takes place between adsorbed CO molecules and

20

O atoms, following the Langmuir-Hinshelwood (L-H) mechanism, on all single crystal

21

surfaces. Studies of CO oxidation on “real” catalysts have also been carried out. Huang and

22

Tsai proposed that Cu2O is the more active among three commercial Cu, Cu2O and CuO

23

powder catalysts, in CO oxidation under oxygen lean and rich conditions. They refer this to a 2

. The kinetics of CO oxidation have been studied on copper single crystal surfaces of


Page 2 of 34

Page 3 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

1

high mobility of lattice oxygen 13. In contrast, Jernigan and Somorjai, using a batch reactor to

2

measure the activity of thin evaporated films, concluded that metallic Cu is the most active

3

phase according to apparent activation energies

4

metallic Cu and Cu2O involves adsorbed O atoms and CO molecules

5

lattice oxygen participates in the reaction and later gets replenished from the gas phase in a

6

redox type mechanism 12,15. The diverging conclusions arrived at from these studies are likely

7

deriving, at least partly, from the choice of catalyst. Although the “real“ powder catalysts may

8

be more relevant for studies aimed at optimizing catalyst formulations, the flat model catalysts

9

offers a more homogeneous system with better control over process conditions making them

10

12

. They proposed that CO oxidation over 12,16

, while over CuO,

better suited for fundamental studies.

11

None of the work mentioned above applied in situ techniques to investigate the oxidation

12

state of Cu catalysts and direct measurements of the reaction rate under steady state conditions.

13

However, it is well established that both the oxidation states and reactivity of catalysts depend

14

on the gas atmosphere. Therefore, it is difficult to use post-reaction measurements to study the

15

correlation between the oxidation state and reactivity. Reversible changes of the catalysts, to

16

states that are only present during use can be missed, thereby introducing errors in the

17

measurement interpretation. Hence, in situ techniques are of great necessity for reliable studies.

18

Recent studies by Xu et al. and Eren et al. used in situ techniques to study Cu single crystal

19

properties under different CO-oxidation conditions 1,8,20. These studies elucidate the sensitive

20

correlation between gas composition, Cu-oxidation state and morphology, but provide

21

estimates of the catalyst activity only via indirect measurements.

22

In this work we have used in situ UV-Vis and mass spectroscopy to investigate the correlation

23

between oxidation states of flat model Cu catalysts and reactivity during carbon monoxide

24

oxidation. Semi in situ X-ray Photoelectron Spectroscopy (XPS) was applied to complement

25

the in situ study. A set of experiments were performed and we found that metallic Cu is the 3


ACS Catalysis

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

1

most active phase in carbon monoxide oxidation but that the metallic phase is only stable under

2

conditions where close to full oxygen conversion is achieved.

3 4

2. EXPERIMENTS

5

2.1 Sample preparation

6

The Cu model catalysts were prepared by evaporation of Cu metal (99.99% pure Cu, Kurt J.

7

Lesker) onto 1.2 cm x 0.9 cm x 0.05 cm quartz wafers in a home-built, stainless steel thermal

8

evaporator. All wafers were calcined at 500 ⁰C overnight in an oven to remove contaminants

9

from the surface prior to Cu-deposition. The nominal thickness of the Cu-film was 5 nm for all

10

samples as measured by an in situ quartz crystal micro balance (Q-pod Quartz Monitor,

11

Inficon). This means that each sample contains ~5 μg of Cu.

12

The Cu/quartz wafer samples were heated to 400 ⁰C, (10 ⁰C/min) and held at that temperature

13

for 1.5 h, first in 200 ml/min, 2.5 vol.% O2/Ar and subsequently in 5 vol.% H2/Ar prior to all

14

measurements, in order to form sintered, stable Cu nanoparticles.

15 16

2.2 Pocket micro-reactor

17

The optical and reactivity measurements were measured simultaneously in a modified

18

commercial instrument, Insplorion X1 (Insplorion AB, Gothenburg Sweden) and a mass

19

spectrometer Pfeiffer Vacuum D-35614, PrismaPlus QME 220. As illustrated in figure 1, the

20

instrument is composed of a quartz tube flow reactor equipped with two parallel sets of optical

21

fibers, which transmits light beams in the wavelength range of 380-1000 nm through the sample

22

holder, to the optical detector. The gas flow through the quartz tube is regulated by mass flow

23

controllers (0-300 ml/min) on the inlet side and the majority of the gas passes through to the

24

exhaust, without being exposed to the catalyst. 4


Page 4 of 34

Page 5 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

1

2 3

Figure 1. Illustration of the micro reactor with optical spectroscopy facilities and sample

4

holder design.

5

The sample holder, consisting of a quartz pocket with the inner dimensions 1.5 cm x 1 cm x

6

0.1 cm is placed inside the quartz tube. When a sample is inserted, a 0.05 cm thin slit is left

7

between the sample surface and the sample holder top wall, forcing the gas to pass in close

8

vicinity to the catalyst material. Due to the sample holder geometry and small volume (the

9

volume inside the sample holder is 0.05 cm3), we refer to this flow reactor as a pocket micro-

10

reactor. The gas enters the sample holder pocket through the open short-end (through which

11

the sample is inserted) and exits through a thin tube (0.1 cm inner diameter, 10 cm length), at

12

the opposite, closed short-end. A capillary restriction is connected by the end of the tube,

13

providing a constant, small leak between the micro reactor and a mass spectrometer. The flow

14

of reactant gases, through the sample holder pocket, ranging from 0.1-10 ml/min (SV, 2 - 200

15

min-1), is regulated by mass flow controllers at the gas outlet. This design promotes sensitivity

16

of measurements and is beneficial for reactivity measurement from small amounts of catalysts.

17

Temperature was measured by a quartz-enclosed thermocouple next to the pocket. The micro

18

reactor combines in situ UV-Vis and mass spectrometry, which provides detailed information

19

about the catalyst oxidation state and reactivity during a reaction.

20 5


ACS Catalysis

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

1

2.3 Optical and reactivity measurement

2

Firstly, we demonstrate the simultaneous UV-Vis and mass spectroscopy measurements

3

during CO oxidation experiments over Cu catalysts in different oxidation states. O2 or H2

4

pretreatments were carried out to oxidize or reduce the Cu catalysts. The samples were heated

5

up in a 5 vol.% O2/Ar or 5 vol.% H2/Ar gas flow to 400 ⁰C at a rate of 10 ⁰C/min, kept at this

6

temperature for 1h and subsequently cooled down. At 50 ⁰C, the oxidizing/reducing gas was

7

purged out and a 6.6 vol.% CO/3.3 vol.% O2 gas flow in Ar was switched on. 60 min at 50 ⁰C

8

was allowed to collect a baseline at stable conditions before the samples were heated up in the

9

reactant gases, at a rate of 10 ⁰C/min to 400 ⁰C. This pre-oxidation/reduction - reaction cycle

10

was repeated three times to assure reproducibility of the results. The reaction flow (through the

11

sample holder pocket) was kept at 1 ml/min.

12

Secondly, we introduce an approach to interpret the optical spectra of the Cu catalyst during

13

a reduction process, from Cu0 to Cu+ and Cu2+, in real time. The catalyst was heated to 400 ⁰C

14

at a rate of 10 ⁰C/min, in a 5 vol.% H2/Ar gas flow, and kept at this temperature for 1h. It was

15

subsequently cooled down, in a 40 vol.% CO/Ar gas flow, to 300 ⁰C where O2 gas was

16

introduced into to the reactor, in small, stepwise increments. Finally, all CO gas was removed

17

from the gas flow, to fully oxidize the catalyst. During this experiment, the reaction flow was

18

kept at 5 ml/min.

19

Finally, addressing kinetic measurements, the samples were heated up in a 5 vol.% H2/Ar

20

gas flow to 400 ⁰C at a rate of 10 ⁰C/min and kept at this temperature for 1h. It was followed

21

by cooling down in the H2/Ar gas flow to reaction temperatures (80, 150 or 250 ⁰C). A 40

22

vol.% CO/Ar flow was switched on and O2 gas was added, in small stepwise increments. In

23

this case, a constant flow of He, 3 ml/min, was added into the gas flow during the whole

24

measurement process. To compensate for signal drift, the MS readouts for O2 and CO2 were 6


Page 6 of 34

Page 7 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

1

normalized to the He-reference. The CO2 signal was further corrected by subtracting reference

2

data from an identical measurement where no catalyst was present in the reactor. The reaction

3

flow was 1 ml/min.

4 5

2.4 X-ray Photoelectron Spectroscopy

6

In order to study the oxidation state and particle dispersion of Cu catalysts, semi in situ XPS

7

experiments are carried out with a Kratos AXIS Ultra spectrometer, equipped with a

8

monochromatic Al Kα x-ray source (Al Kα= 1486.6 eV). The measurements were performed

9

at a 40 eV pass energy with a 0.1 eV step size and the background pressure in the chamber was

10

typically lower than 10-8 mbar.

11

Reactions were performed in a high-temperature gas reaction cell (Kratos Analytical Ltd,

12

Manchester, U.K.), which consists of a dome-shaped quartz chamber, allowing for in vacuo

13

transfer into the measurement chamber. A quartz stub was used as sample support and the

14

chamber pressure was kept at 1.5 bar during all treatments. Initial oxidation treatments at 400

15

⁰C, were performed in a 10 vol.% O2/Ar gas flow, to fully oxidize the catalyst and remove

16

adventitious carbon prior to experiments. For the O2, H2 and CO/O2 treatments (presented in

17

3.1) the sample was heated up to 400 ⁰C at a rate of 10 ⁰C/min and kept at this temperature for

18

1h, followed by cooling down to 50 ⁰C. For the experiment investigating oxidation of a pre-

19

reduced catalyst under stoichiometric reaction conditions (presented in 3.2), the sample was

20

pre-reduced in H2 at 400 ⁰C for 1 h followed by cooling down to 50 ⁰C. Then a 6.6 vol.%

21

CO/3.3 vol.% O2 gas flow in Ar was was introduced. The sample was heated to the indicated

22

temperatures. At the targeted temperature the gas flow was switched off and the reaction cell

23

was immediately evacuated, rapidly cooled down and transferred into the measurement

24

chamber. 7


ACS Catalysis

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 34

1

The results were analyzed with Casa XPS software (version 2.3.16 Pre-rel 1.6) and all

2

spectra were calibrated using the Si 2s peak at 154.4 eV. For quantitative analysis, the Si 2s,

3

Cu 2p and O 1s and the Wagner relative sensitivity factors were used. In order to distinguish

4

Cu0 from Cu+, the modified Auger parameter was calculated according to equation (1):

5 6

Equation 1

α = hν + KE (Cu LMM) - KE ( Cu 2p3/2)

7 8

where hν is the X-ray photon energy, KE(Cu LMM) and KE(Cu 2p3/2) is the kinetic energy of the

9

LMM Auger electron and the 2p3/2 photoelectron, respectively.

10 11 12 13

2.5 SEM Scanning electron microscopy (SEM) was carried out in a FEI Quanta 3D FEG dual beam instrument to obtain morphology information of Cu nanoparticles.

14 15

3 RESULTS

16

3.1 Cu, Cu2O and CuO, XPS and UV-vis reference spectra

17

In order to obtain reference spectra of Cu, Cu2O and CuO, UV-vis measurements were

18

performed after treatments at three well defined conditions. Inspired by the work by Jernigan

19

and Somorjai 12, we used pre-treatments at 400 ⁰C for 1 h, in three different gas environments

20

(5 vol.% H2, 9 vol.% CO/1 vol.% O2, and 5 vol.% O2 in Ar) to convert the 5 nm thick,

21

evaporated Cu films into metallic Cu(0), Cu+ and Cu2+-oxides, respectively. Semi in situ XPS

22

analysis was then applied to verify the resulting oxidation states. Figure 2(a, b) displays the Cu

23

2p and O 1s spectra and the analysis details are shown in table 1 together with reference values

24

from the literature 9,11,21–26. 8


Page 9 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

1

2 3 4

Figure 2 (a) Cu 2p XPS spectra, (b) O 1s and Cu LMM spectra and (c) UV-Vis spectra of Cu catalysts after gas treatments at 400 ⁰C resulting in Cu(0), Cu+ and Cu2+ respectively. 9


ACS Catalysis

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

1

Page 10 of 34

The satellite peaks present in the Cu 2p spectrum of the O2-pretreated Cu, confirms the 10

2

formation of CuO

. Furthermore, the peak at 529.8 eV observed in the O1s region,

3

corresponds to lattice oxygen of CuO 11,21 while the larger peak at 532.5 eV corresponds to the

4

lattice oxygen of SiO2 27. The molar O/Cu ratio is 0.81, close to but slightly below what is

5

expected for CuO. This is likely due to that part of the O is bound to surface defect sites, which

6

are inherent to the CuO phase. This gives rise to a photoelectron peak at a higher binding energy

7

and has been reported in the literature for many pure oxides, including copper, nickel and

8

chromium 22,28,29. The defect related peak can comprise as much as 20 - 40 % of the main peak

9

but due to the higher binding energy it is overlapping with the large O peak related to SiO2 22,

10

thus explaining the low O/Cu ratio. It should be kept in mind that these surface defect sites are

11

likely to contribute to the catalytic activity of the oxide phase. In contrast, no satellite peaks

12

appear in the spectra of the catalyst treated in CO/O2 gas mixture. The Cu 2p3/2 peak appears at

13

932.4 eV and the modified Auger parameter is 1849.2 eV. Simultaneously, the lattice oxygen

14

peak, corresponding to copper oxide, shifts to 530.4 eV and the atomic O/Cu ratio decreases to

15

0.48, confirming the formation of Cu2O 9,11,21–23. Metallic Cu is formed after H2-pretreatment

16

according to the modified Auger parameter (1851 eV) and the absence of a peak corresponding

17

to lattice oxygen of copper oxides. The Cu 2p peak appears at a slightly lower binding energy

18

than expected after the H2-treatment. This is likely due to a difference in charge build-up in the

19

SiO2 surface layer when the metallic overlayer is present 30. Another interesting observation is

20

that the O 1s peak at 532.5 eV, corresponding to lattice oxygen in SiO2, is observed after all

21

three treatments, suggesting that the Cu film does not form a continuous layer after any of the

22

three pre-treatments, but rather a discontinuous particle film. Moreover, the Si/Cu ratio changes

23

according to the oxidation states after different treatments, from 1.13 – 1.31 for the copper

24

oxides to 4.85 for the metallic Cu. This indicates significant particle reshaping upon oxidation

25

and reduction. 10


Page 11 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

1 2

Table 1. XPS information of Cu catalysts after different treatments from this work and from

3

literature references

Samples

T (⁰C)

BE Cu 2p 3/2 (eV)

Modified Auger parameter (eV)

BE O1s (eV)

O/Cub

Si/Cuc

O2 treatment

400

933.8

Sa

529.8

0.81

1.31

H2 treatment

400

932.1

1851

-

0

4.85

CO/O2 treatment

400

932.4

1849.2

530.4

0.48

1.13

50

932.0

1849

530.1

0.46

5.51

200

932.5

1849

530.3

0.54

2.13

300

933.8

Sa

529.8

0.81

1.35

400

933.4

Sa

529.7

0.79

1.14

933.4-933.8

Sa

529.5-529.7 11,21

-

9,22,23

1849.2-1849.6 9,23

530.2 11,21,23

-

932.5-932.7

1851.2-1851.6

-

-

Reduced Cu in 6.6 vol.% CO/3.3 vol.% O2

4 5 6

CuO

-

Cu2O

-

Cu

-

9,22–26

932.2-932.4

9,22–26

9,22,24–26

a. S denotes the existence of satellite peaks of CuO b, c. The Si 2s, Cu 2p and O 1s and the Wagner relative sensitivity factors peaks were used to calculate atomic O/Cu and Si/Cu ratios.

7 8

The XPS results show that Cu, Cu2O or CuO are obtained after H2, CO/O2 and O2

9

pretreatments, respectively. Therefore, these three treatment conditions were applied to get

10

UV-Vis extinction reference spectra of Cu, Cu2O and CuO, which are displayed in Figure 2(c).

11

Each pretreatment was applied for 1h although the optical spectra were stable already after the 11


ACS Catalysis

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 34

1

first 15 min. This indicates that the phase transformation was readily completed during the pre-

2

treatment and not hampered by slow diffusion limited processes. After H2-treatment a clear

3

peak at around 580 nm appears in the spectrum. This peak stems from the localized surface

4

plasmon resonance (LSPR), which is a signature of metallic Cu nanoparticles

5

optical patterns of CuO and Cu2O are lacking the LSPR peak but are mutually distinguishable

6

by the difference in extinction between 400 and 700 nm. Thus the optical spectra of Cu

7

nanoparticles in these three different oxidation states are used as references in the following

8

analysis.

7,31,32

. The two

9

In addition to the chemical and optical differences between the evaporated thin Cu-films

10

after different gas treatments, significant morphological differences are observed. Figure 3

11

shows SEM images of the 5 nm Cu-films after heating for extended time at 400 ⁰C in Ar-

12

diluted O2, H2 and 6.6 vol.% CO/3.3 vol.% O2 mixtures.

13 14

12


Page 13 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

1 2

Figure 3. SEM-images of a 5 nm thick Cu-film after sequential treatment at 400 oC in (a) H2

3

1h, O2 1h and 6.6 vol.% CO/3.3 vol.% O2 1h (oxidized) and (b) H2 1h, O2 1h, 6.6 vol.% CO/3.3

4

vol.% O2 1h and H2 1h (reduced).

5

From these images, it is clear that the Cu-film is discontinuous and exhibits different

6

morphologies after oxidizing and reducing treatments. The oxidizing treatment (figure 3a)

7

leads to the formation of a network of flat nanoparticles. This corresponds well with the XPS

8

results where a smaller Si/Cu ratio was seen in this case. In contrast, the reducing treatment

9

leads to clearly separated nanoparticles with a rounder shape, which is responsible for the larger 13


ACS Catalysis

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

1

Si/Cu ratio as measured by XPS. The average diameter of the metallic Cu-nanoparticles in

2

figure 3b is estimated to 31±12 nm.

Page 14 of 34

3 4

3.2 CO oxidation on reduced and oxidized Cu in CO/O2

5

In order to demonstrate the power of simultaneous UV-Vis and mass spectroscopy

6

measurements, CO oxidation experiments were carried out in 6.6 vol.% CO/3.3 vol.% O2 in

7

Ar, over a pre-oxidized and a pre-reduced Cu catalyst. The pre-oxidized sample is expected to

8

remain as CuO in the stoichiometric gas composition, while the reduced, metallic Cu is

9

expected to oxidize during the experiment 12. Figure 4 displays CO2 formation from the MS

10

and the UV-Vis spectra, taken during the experiments (solid lines) as well as the reference

11

spectra from figure 2 (dash-dotted lines).

14


Page 15 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

1 2

Figure 4. (a) Mass spectrometer measurements of CO2-formation over pre-oxidized and pre-

3

reduced Cu catalyst, and in situ UV-Vis spectra of (b) a pre-oxidized and (c) a pre-reduced Cu

4

catalysts during CO oxidation in Ar-diluted 6.6 vol.% CO/ 3.3 vol.% O2 mixture. The

5

temperature was increased by 10 ⁰C/min to 400 ⁰C, where it was kept for 200 min. 15


ACS Catalysis

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 34

1

Figure 4(a) shows the CO2 evolution from a pre-reduced and a pre-oxidized Cu sample

2

during temperature ramps and subsequent isothermal reaction at 400 ⁰C in the CO/O2 gas

3

mixture. The rate of CO2 formation increases smoothly with increasing temperature and both

4

samples show significant catalytic activity at temperatures above 200 ⁰C. However, the rate of

5

CO2 formation is lower over the pre-reduced than over the pre-oxidized catalyst during the

6

entire experiment. During the extended, isothermal reaction at 400 ⁰C the CO2 formation of the

7

pre-oxidized catalyst is fairly stable, exhibiting only a slight decrease in activity. In contrast,

8

the isothermal reactivity of the pre-reduced catalyst initially decreases slightly and then

9

increases with time. After running the reaction for 200 min, it appears that the reactivity still

10

approaches that of the pre-oxidized catalyst, indicating that the catalyst adaptation to the gas

11

environment is rather slow, even at 400 ⁰C.

12

In order to correlate the changes in reactivity with the Cu oxidation state, Figure 4(b) shows

13

the optical spectra of the pre-oxidized catalyst at selected moments during the CO-oxidation

14

experiment. All spectra resemble that of CuO and no significant changes are observed,

15

indicating that the CuO is stable under these reaction conditions. Only after extended reaction

16

at 400 oC, a slight shift in the spectral intensity at the shorter wavelengths can be observed,

17

suggesting a minor, partial reduction of the CuO to Cu2O.

18

In figure 4(c), the corresponding UV-Vis spectra for the pre-reduced catalyst are shown.

19

These spectra exhibit a number of interesting changes during the CO-oxidation experiment.

20

The first thing to notice is the LSPR-peak at around 580 nm after H2-reduction, which confirms

21

the presence of metallic Cu nanoparticles

22

longer wavelengths already at 50 ⁰C when the H2/Ar gas mixture is replaced with CO/O2,

23

implying formation of a surface oxide on the Cu nanoparticles

24

increased, the peak is broadened and shifts further to longer wavelengths. At 175 ⁰C the LSPR-

25

peak is completely gone, demonstrating that the Cu nanoparticles are further oxidized. As

7,31,32

. The LSPR-peak position shifts slightly to

16


7,31,32

. As temperature is

Page 17 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

1

temperature is increased to 400 ⁰C, the optical extinction at wavelengths shorter than 500 nm

2

increases, thereby resembling the reference spectra of Cu+. During the extended period at 400

3

⁰C, the spectral intensity at the wavelength between 500 and 700 nm gradually increases,

4

eventually resembling the reference spectra of Cu2+.

5

To complement the in situ UV-Vis measurements, semi-in situ XPS was applied to study

6

the pre-reduced catalyst after exposure to the same, oxidizing reaction conditions. Figure 5

7

shows the Cu 2p and O 1s peaks of the pre-reduced catalyst after exposure to the CO/O2 gas at

8

various temperatures as well as the reference spectra of a pre-reduced and a pre-oxidized

9

catalyst. In addition, the peak positions, modified Auger parameter and the O/Cu ratios are

10

presented in table 1.

11 12

17


ACS Catalysis

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 34

1 2

Figure 5. Semi in situ XPS spectra of the (a) Cu 2p and (b) O 1s region of the pre-reduced Cu

3

catalyst after treatment in the 6.6 vol.% CO/3.3 vol.% O2 gas mixture at different temperatures.

4

After switching from H2 to the CO/O2 gas mixture at 50 ⁰C, a peak in the O1s spectra (figure 11,21,23

5

5b) at 530.1 eV appears, corresponding to oxygen in cuprous oxide

6

modified Auger parameter changes from 1851 to 1849 eV and the O/Cu molar ratio is 0.46.

7

This indicates oxidation of the surface of metallic Cu into Cu2O 9,23, which confirms that the

8

red-shift of the peak in the UV-Vis spectra is due to the formation of an oxide shell around the

9

metallic Cu nanoparticles. As temperature increases to 200 ⁰C, the Cu 2p3/2 peak shifts to 932.5

10

eV, while the modified Auger parameter and O1s peak position of Cu oxide remains stable,

11

indicating that the oxidation state of the catalyst does not change. The O/Cu molar ratio 18


. Furthermore, the

Page 19 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

1

increases to 0.54, which indicates a slight further oxidation. At 300 ⁰C, satellite peaks appear

2

in the Cu 2p spectra and the oxygen peak corresponding to copper oxide shifts to 529.8 eV. At

3

the same time, the atomic O/Cu ratio increases to around 0.81, indicating that the catalyst

4

surface was further oxidized into CuO. Increasing the temperature to 400 ⁰C leads to no further

5

changes in the spectra. During the reaction process, it is also noted that the Si/Cu ratio (table

6

1) steadily decreases as temperature is increased. This implies that Cu nanoparticles wets the

7

Si surface, thereby covering a larger fraction of the surface.

8

The amount of C on the catalyst surface (not shown) is around a few percent after all

9

treatments, which shows that the pre-treatments effectively remove adventitious carbon. In

10

particular, no increase is seen in the C 1s signal after the CO/O2 treatments, indicating that no

11

extra carbon species are introduced that can obstruct the subsequent reactions.

12

In summary, the pre-reduced catalyst forms a surface oxide upon exposure to the 6.6 vol.%

13

CO/ 3.3 vol.% O2 gas mixture, leading to the formation of a core-shell structure. With

14

increasing temperature, the oxide shell becomes thicker and the metallic Cu core smaller until

15

the catalyst is completely oxidized. During the isothermal reaction at 400 ⁰C, the oxidized Cu

16

is further oxidized into CuO whereupon the nanoparticles wet the surface, thereby exposing a

17

larger surface area. The pre-reduced catalyst is thus exposing a surface with mixed oxidation

18

states during the ramp experiment while the pre-oxidized catalyst is completely CuO

19

terminated.

20 21

3.3 Generalized interpretation of UV-Vis spectra

22

As demonstrated in the previous section, the optical spectra collected during the catalytic

23

experiments provide a valuable link between the catalyst activity and oxidation state. Optical

24

spectra can be obtained at a time resolution down to a few milliseconds thus generating a vast 19


ACS Catalysis

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 34

1

amount of data. It is therefore desirable to introduce a generalized interpretation of the optical

2

spectra to indicate when significant changes take place during an experiment. When LSPRs are

3

used in sensing applications, the peak position and/or peak width is frequently used to achieve

4

a convenient readout of optical changes

5

an approach would only be of limited use, since the plasmonic peak is completely eliminated

6

upon full oxidation of the nanoparticles. We therefore choose to calculate the “optical center”,

7

defined as the wavelength at which the integrated extinction spectra equals half of the total

8

integral value:

32–35

. For the experiments demonstrated above, such

9 10

𝝀

𝟏

𝝀𝒎𝒂𝒙

∫𝝀𝒎𝒊𝒏 𝑬(𝝀)𝒅𝝀=𝟐 ∫𝝀𝒎𝒊𝒏 𝑬(𝝀)𝒅𝝀

Equation 2

11 12

in which λ is the wavelength and E(λ) is the corresponding intensity of the extinction spectra.

13

This approach allows us to detect minor changes in the optical spectra and to conveniently

14

follow optical changes over time, both from metallic to oxidized Cu and from one oxide to

15

another. The usefulness of this approach is demonstrated by an experiment where a pre-reduced

16

Cu catalyst is exposed to a constant flow of CO, 40 vol.% diluted in Ar, with a gradually

17

increasing O2-content in the gas flow at 300 ⁰C, thereby inducing sequential oxidation of

18

metallic Cu to Cu2O and eventually CuO. To assure a uniform response over the sample, a high

19

flow rate (5 ml/min) and thus low rates of O2 conversion was applied. Figure 6 (a, b) shows the

20

calculated optical center and selected UV-Vis spectra during the experiment, plotted versus

21

time (bottom x-axis) and the relative oxygen concentration (top x-axis).

22

20


Page 21 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

1 2

Figure 6. (a) Optical center position of the in situ UV-Vis spectra during phase transitions

3

from Cu0 to Cu2O and to CuO in O2/CO gas mixtures with increasing oxygen content at 300

4

⁰C, (b) Selected optical spectra from points where the optical center shows large shifts.

5

The optical center position displayed in figure 6(a) shows major shifts in two regions, i)

6

O2/CO ratios between 0.95×10-2 and 1.25×10-2 and ii) higher than 4.5×10-2. By looking closer

7

into the optical spectra in these regions the nature of these shifts can be clarified.

8

The large shift of the optical center for the O2/CO ratios between 0.95×10-2 and 1.05×10-2

9

derives from a shift towards longer wavelengths and a broadening of the Cu nanoparticle

10

LSPR-peak in Fig 6(b). This LSPR-shift stems from the formation of a Cu2O shell around the 21


ACS Catalysis

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 34

7,32,36,37

1

Cu nanoparticles, as demonstrated in several previous studies

. For O2/CO ratios

2

between 1.1×10-2 and 1.25×10-2, the optical center shifts back towards shorter wavelengths

3

again. The optical spectra from this part of the experiment, shown in Fig 6(b) reveals that the

4

plasmonic peak is completely eliminated while the extinction at wavelengths between 400 and

5

500 nm increases, eventually resembling the Cu+ reference spectra. This indicates that the

6

surface oxidation has reached all the way into the Cu-particle cores. The first, sharp peak in the

7

optical center plot can thus be associated with oxidation of the metallic Cu nanoparticles to

8

Cu+.

9

The second, large shift in the spectra, observed for O2/CO ratios larger than 4.5×10-2 can

10

then be associated with oxidation of the Cu+ to Cu2+. The optical spectra reveals decreasing

11

extinction between 400 and 500 nm and increasing extinction for wavelengths between 500

12

and 700 nm, gradually transforming the optical spectra from that of the Cu+ reference to one

13

closely resembling the Cu2+.

14

The third and more subtle shift in the optical center plot can be observed at the point where

15

oxygen is first introduced into the CO gas stream. This initial, tiny shift in the LSPR peak

16

position would not be detectable without the mathematical analysis. It can tentatively be

17

attributed to O replacing CO as the majority species adsorbed on the surface of the Cu

18

nanoparticles. Similar effects have previously been seen on Pt nanoparticles supported on silica

19

embedded Au discs 33 and during water adsorption/desorption from Au-particles 38.

20

In summary, the optical center analysis enables us to monitor the oxidation of Cu

21

nanoparticles in real time and thus to compare the stability of the Cu oxidation states under

22

different conditions.

23

22


Page 23 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

1

ACS Catalysis

3.4. Kinetic measurements of CO oxidation on metallic Cu

2

The results presented in section 3.2 and 3.3 shows that our micro reactor can be used to

3

measure the reactivity of a flat model catalyst during CO-oxidation, while simultaneously

4

monitoring the oxidation state. We therefore decided to study the kinetics of CO oxidation on

5

metallic Cu, which shows fairly contradictory results in the literature

6

Experiments to measure O2 reaction orders in CO oxidation were performed in a constant CO-

7

flow (40 vol.% in Ar), with various O2/CO molar ratios, first increasing stepwise from 0 to

8

4×10-2 in intervals of 0.5×10-2 step-1 and then decreasing back to 0 in equivalent steps. Figure

9

7 shows the in situ O2 and CO2 MS signals from experiments at three different temperatures

10

and the corresponding optical center plots of the UV-Vis spectra. These temperatures were

11

chosen to investigate the stability of the catalyst at different levels of oxygen conversion, at

12

temperatures up to or below where Cu-catalysts are typically used.

13 14 15 16

23


6,8,12,13,17–19,39

.

ACS Catalysis

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 34

1 2

Figure 7. Mass spectrometer measurement of (a) O2 content and (b) CO2-formation during CO

3

oxidation on pre-reduced flat model Cu catalysts at 80, 150 and 250 ⁰C at constant CO-flow

4

and various O2/CO ratios. The dotted curve in (a) shows the oxygen content when no catalyst

5

is present, i.e. with no conversion. (c) Optical center position (indicating the observed LSPR-

6

shifts), calculated from the in situ UV-Vis measurements during the CO-oxidation experiments.

7

The dotted curve in figure 7(a) corresponds to a blank measurement without catalyst and

8

serves as a reference to show the amount of O2 in the gas flow when no reaction takes place.

9

At 250 ⁰C all of the O2 is consumed for O2/CO ratios below 3×10-2 while at O2/CO = 3×10-2

10

roughly half of the O2 is left in the gas flow after passing the catalyst. For higher O2/CO ratios,

11

the conversion is negligible in comparison, implying a catalyst deactivation. It is also

12

interesting to point out that O2-conversion is not regained when the oxygen concentration is

13

decreased again. A similar behavior is observed at 150 ⁰C although full O2-conversion is only 24


Page 25 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

1

achieved for the two lowest O2/CO ratios and catalyst deactivation is observed already at

2

O2/CO=1.5×10-2. Compared with the reactions at high temperature, roughly one fourth of the

3

O2 is left in the gas flow at the first step and almost no O2 consumption is observed for a gas

4

stream with the O2/CO ratios higher than 1×10-2 at 80 ⁰C.

5

Corresponding changes are also observed in the CO2 curves in figure 7(b). Initially, the

6

amount of CO2 produced is increased with increasing the O2/CO ratios, while the CO2

7

production decreases drastically for higher O2/CO ratios. No reactivity is recovered during the

8

O2 decrease processes at all three temperatures, which indicates that an irreversible catalyst

9

deactivation has occurred. This set of experiments shows that the reactivity of the initially

10

metallic Cu-catalysts gradually declines and eventually deactivates as the O2-partial pressure

11

is increased beyond the point where full O2-conversion is achieved and that it is an irreversible

12

process.

13

The in situ UV-Vis spectroscopy can now be applied to monitor the oxidation states of the

14

Cu catalysts during the kinetic measurements. Figure 7(c) shows the shifts in the optical center

15

for the reactions at three temperatures. An initial, slight shift to longer wavelengths is seen at

16

all three temperatures when O2 is introduced into the gas flow, tentatively assigned to a shift in

17

the Cu LPSR peak position when the CO molecules adsorbed on the Cu surface is replaced by

18

O 7,31–33. However, the most important observation from these curves is that large shifts in the

19

optical center position correlate in time with the observed decrease in catalyst activity.

20

According to the results presented in 3.2 and 3.3, these shifts indicate oxidation of the metallic

21

Cu to Cu2O.

22

It is also interesting to note that the oxide shell continues to grow, rather than reduce back

23

to metallic Cu, even when the O2-concentration is decreased again. At 250 ⁰C, the Cu-oxidation

24

is suppressed up to a relatively high oxygen concentration but proceeds quickly to Cu2O once

25

it has begun. At 80 ⁰C, in contrast, oxidation starts already at the O2/CO ratio 1×10-2, but does 25


ACS Catalysis

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 26 of 34

1

not proceed to completely oxidized Cu2O, while the experiment at 150 ⁰C shows an

2

intermediate behavior. Finally, it should be pointed out that no re-reduction of the Cu catalysts

3

takes place, as long as O2 is present in the gas flow. However, when CO was the only reactive

4

gas in the flow, a rapid re-reduction was observed for the experiments at 150 and 250 ⁰C, as

5

seen by the sharp spikes towards the end of the optical center plots, while at 80 ⁰C, the catalyst

6

stays oxidized.

7 8

4. DISCUSSION

9

The newly developed micro reactor facilitates simultaneous measurements of reactivity and

10

oxidation states of flat model catalyst using a combination of in situ MS and UV-Vis

11

spectroscopy. This is especially powerful for catalysts consisting of metallic Cu nanoparticles,

12

since these exhibit localized surface plasmon resonances (LSPR). The spectral position of these

13

LSPR-peaks is extremely sensitive to the formation of surface oxides, which has been

14

demonstrated in many previous publications

15

exceptionally well suited to study how the oxidation of metallic Cu catalysts changes the

16

catalytic activity. By mathematical integration of the optical spectra, it is possible to

17

continuously monitor very subtle changes in the catalyst state. Analysis of the integrated optical

18

spectra allows us to follow, not only oxidation of Cu 0 to Cu2O (through shifts in the LSPR

19

peak position) but also the oxidation of Cu2O to CuO in real time.

7,31,32,35

. This experimental setup is therefore

20

The activity during temperature-programmed surface reaction (TPSR) experiment from 50

21

to 400 ⁰C in a stoichiometric gas composition (CO/O2=6.6/3.3), is higher on a pre-oxidized

22

than on a pre-reduced Cu catalyst, seemingly suggesting that CuO is more active than metallic

23

Cu. However, the oxidation state of the pre-oxidized Cu catalyst remained stable as Cu 2+ under

24

the entire experiment. In contrast, the pre-reduced Cu catalyst gradually oxidized during the 26


Page 27 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

1

reaction, as evidenced by the in situ UV-Vis and supporting semi in situ XPS. Already at 50

2

⁰C the metallic Cu surface was oxidized into Cu2O by the reactant gases, eventually forming

3

fully oxidized CuO. This experiment thus rather shows that CuO is more active than the surface

4

oxidized Cu catalyst, already at low temperatures.

5

The oxidation of metallic Cu to Cu2O at 50 ⁰C is confirmed by the semi in situ XPS

6

measurements. However, the XPS-measurements indicate that CuO is formed already during

7

the temperature ramp (at T ≥ 300 ⁰C), while the UV-Vis measurements suggest that CuO

8

formation takes place mainly during the extended reaction at 400 ⁰C. At this point it is not clear

9

if this discrepancy stems from the different geometry and flow conditions in the micro reactor

10

and the XPS-reaction cell or if a slight CuO surface oxidation can remain undetected in the

11

UV-Vis spectra. In any case, it is clear that the surface of the pre-reduced catalyst is

12

transformed into CuO, already during the temperature ramp or the early stages of reaction at

13

400 ⁰C, yet the reactivity of the pre-reduced catalyst remains lower than the pre-oxidized.

14

It is assumed that the reaction on CuO takes place between adsorbed CO molecules and

15

lattice oxygen of CuO, which is replenished by gas phase O2 12,14,15. Therefore, the amount of

16

lattice oxygen of CuO should be related to its reactivity. In other words, more lattice oxygen

17

from CuO is available for CO oxidation in the pre-oxidized Cu catalyst than from the CuO

18

shell-Cu2O core of the pre-reduced Cu catalyst. This suggests that sub-surface oxygen is

19

important for the reaction rate, as demonstrated in previous research 6. An additional

20

explanation to the observed differences in reactivity could be that the nanoparticle shape (SEM)

21

and hence the exposed surface area (XPS) is different for the pre-reduced and the pre-oxidized

22

samples and that reshaping of the CuO nanoparticles proceed slowly during the reaction

23

conditions.

24

In section 3.4, isothermal, kinetic measurements of CO oxidation on metallic Cu were

25

carried out at three different temperatures. Gradual deactivation of metallic Cu was observed 27


ACS Catalysis

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 28 of 34

1

with increasing O2 partial pressure at all temperatures, coinciding with oxidation of the catalyst.

2

It is well accepted that the reaction order in O2 is negative over metallic Cu, which is due to

3

that absorbed atomic O interacts strongly with the metal surface and blocks active sites so that

4

CO cannot adsorb 12,17–19. However, lack of knowledge about the oxidation state of the catalyst

5

during the reactions post uncertainties in studies of the reaction mechanism and active phase.

6

Metallic Cu is well known to be sensitive to oxygen content in the gas atmosphere and in situ

7

characterization techniques are therefore of great necessity to indicate whether it remains stable

8

during measurements. Indeed, a recent, near ambient pressure XPS study by Eren and co-

9

workers 8, demonstrates how CO adsorbed on a Cu surface is replaced by O when 3% of O2 is

10

added to the CO gas stream during CO oxidation. Under these conditions, O was adsorbed on

11

the metallic Cu, rather than forming an oxide, while virtually no CO could be observed on the

12

surface. At higher O2 concentrations, they observed both surface and subsurface oxidation to

13

Cu2O. This is consistent with our observations that metallic Cu is stable, only as long as near

14

to full O2 conversion is achieved. During the reaction, free surface sites will continuously be

15

formed upon desorption of the products, allowing new reactants to adsorb. At near full O2

16

conversion, only CO is available to adsorb on these free sites thus sustaining the reaction. At a

17

lower rate of O2 consumption, the free surface sites will instead be occupied by O, blocking

18

the adsorption of CO, thereby hindering the CO2 formation. The adsorbed oxygen will then

19

instead form an oxide with the Cu. The rate of Cu-oxidation depends strongly on temperature,

20

since it requires diffusion through the bulk of the Cu particles. This can also be seen from our

21

measurements, where Cu oxidation, once it begins, proceeds quickly at 250 ⁰C. In contrast, at

22

80 ⁰C full conversion is not achieved, even for the lowest oxygen contents. Cu oxidation and

23

deactivation thus starts immediately but does not complete during the entire time it is exposed

24

to O2 containing gas (7.5 h).

28


Page 29 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

1

Furthermore, reducing the oxygen content in the CO/O2 gas mixture does not restore the

2

catalyst activity. The stability of metallic Cu requires close to full O2-conversion but the

3

oxidized Cu is significantly less active than the metallic phase. Therefore, once the catalyst

4

surface has been oxidized, the O2-conversion drops significantly and the catalyst remains

5

oxidized even though the gas composition is restored to conditions where the metallic phase

6

was initially stable.

7

Kinetic measurements should be attained at reaction conversion below 20 % to ensure the

8

reaction takes place within the kinetically limited regime, avoiding transport limitations. Our

9

measurements points to a fundamental problem in this respect, since metallic Cu is only stable

10

in low O2 flow and high conversion. In other words, full or nearly full O2 conversion is

11

necessary to ensure the stability of metallic Cu under continuous reaction conditions, which

12

are not suitable for kinetic measurements.

13

Combined, our measurements show that the reactivity of Cu (0) > CuOx. This is in agreement

14

with previous results by Jernigan and Somorjai and by Szanyi and Goodman

15

disagreement with other studies

16

oxidation is likely at the root of these conflicting reports.

8,13

12,19

but in

. The extreme sensitivity of metallic Cu towards surface

17

Also the fundamental difference between powder- and flat model catalyst are likely to

18

contribute to the diverging results reported in the literature. The combination of in situ UV-Vis

19

and in situ catalytic testing has been reported in previous work 40–42. Although these studies do

20

not address the same catalytic system as our study, they are interesting as a comparison for the

21

experimental approach. Smeets et al. and Kondratenko et al. measured characteristic electronic

22

transitions of copper containing powder catalysts in the diffuse reflectance mode (DRS)

23

In the latter work, the Cu LSPR peak was used as indicator to study the catalyst oxidation state

24

under steady-state and transient reaction conditions

42

40,41

.

. However, they did not use any

29


ACS Catalysis

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

1

complementary measurements to identify the characteristic optical spectra corresponding to

2

the different oxidation states.

Page 30 of 34

3

The obvious advantages with using powder catalysts for the studies are that it is easier to

4

perform activity tests and that real, commercial catalysts can be used. However, DRS

5

measurements from a packed bed rector only provide information about the material at the

6

surface of the bed, close to the reactor wall. The reactor wall can act either as a heat sink, for

7

the case of exothermic reactions, or as a heat source when the reactor is externally heated by

8

an external oven 43. Furthermore, catalyst particles at the reactor boundary are surrounded by

9

catalytic material on one side and the reactor wall on the other side. It is therefore likely that

10

the conversion is locally different near the reactor wall compared to at the center of the reactor.

11

These effects can cause radial gradients in the temperature and gas composition. Therefore, it

12

cannot easily be excluded that catalyst material close to the reactor wall in a packed bed reactor

13

is different from the material at the center of the catalyst bed. In contrast, using flat model

14

catalysts in combination with transmission mode UV-Vis spectroscopy has the clear advantage

15

that all the material contributing to the activity is accessible to the optical characterization.

16

Therefore, even if temperature or gas phase gradients results in a catalyst with spatially

17

different properties (e.g. oxidation state) this will not remain undetected. Furthermore, gases

18

are subject to diffusion limitations in the pores of supported powder catalysts 44. Therefore, a

19

faster, more homogeneous response is expected from flat model catalyst, where diffusion

20

limitations, both within the support pores, within the metal particles and in the gas phase are

21

kept at a minimum. In combination with the acquisition of reference spectra taken under

22

conditions where the three different Cu-phases are shown to be stable, our study results in more

23

reliable interpretation of the optical spectra and correlation between the oxidation state and

24

catalytic activity.

25 30


Page 31 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

1 2

6 Conclusion

3

We have developed a new flow reactor, equipped with in situ UV-Vis and mass

4

spectroscopy, capable of measuring the reactivity of flat model catalysts with as little active

5

material as a few μg. This combination makes it a powerful technique to correlate reactivity

6

and oxidation state of the catalyst during reaction conditions, while the flat model catalyst

7

ensures that all catalyst material is exposed to the gas phase reactants under minimum diffusion

8

limitations. CO oxidation on Cu-catalysts have been used to demonstrate that we can get

9

simultaneous information about reactivity and oxidation state from tiny amounts of evaporated

10

Cu model catalysts (~5 μg) during a reaction process.

11

In the CO-oxidation experiment, metallic Cu was found to be the most active phase. Backed

12

up by semi in situ XPS measurements, the in situ UV-Vis allowed us to follow the oxidation

13

of Cu0 to Cu2+, via a core (Cu0) -shell (Cu+) structure, in detail during the CO-oxidation

14

experiment. Moreover, we have introduced the optical center as an analog to the peak position

15

conventionally used to display subtle changes in nano plasmonic sensing. This allows us to

16

follow, not only the phase transition from metallic Cu to Cu+, but also from Cu+ to Cu2+ in real

17

time. Kinetic measurements were attempted on metallic Cu. It was found that catalyst

18

deactivation took place due to an irreversible oxidation of the metallic Cu for oxygen

19

concentrations as low as 0.03 vol.% and below, depending on temperature and the resulting

20

level of oxygen conversion during the reaction. Full O2 conversion is necessary to ensure the

21

stability of metallic Cu, suggesting that kinetic measurements under realistic, steady state

22

conditions are difficult or impossible to perform.

23 24

AUTHOR INFORMATION 31


ACS Catalysis

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

1

Corresponding Author

2

*E-mail: [email protected]

3

Notes

4

The authors declare no competing financial interest.

5

ACKNOWLEDGMENT

6

We gratefully acknowledge funding by the China Scholarship Council for enabling Ms. Bu to

7

perform her doctoral project at the Eindhoven University of Technology. The authors

8

gratefully acknowledge partial financial funding from Synfuels China Technology Co., Ltd.

Page 32 of 34

9 10 11

REFERENCES

12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33

(1)

Eren, B.; Lichtenstein, L.; Wu, C. H.; Bluhm, H.; Somorjai, G. A.; Salmeron, M. J. Phys. Chem. C 2015, 119 , 14669–14674.

(2)

Chen, M. S.; Goodman, D. W. Science. 2004, 306, 252-255.

(3)

Schubert, M. M.; Hackenberg, S.; Van Veen, A. C.; Muhler, M.; Plzak, V.; Behm, R. J. J. Catal. 2001, 197, 113–122.

(4)

Xie, X.; Li, Y.; Liu, Z. Q.; Haruta, M.; Shen, W. Nature 2009, 458, 746–749.

(5)

Yoon, B.; Hakkinen, H; Landman, U.; Worz, A. S.; Antonietti, J. M.; Abbet, S.; Judai, K.; Heiz, U. Science 2015, 45, 435–452.

(6)

Van Pruissen, O. P.; Dings, M. M. M.; Gijzeman, O. L. J. Surf. Sci. 1987, 179, 377-386.

(7)

Marimuthu, A.; Zhang, J.; Linic, S. Science 2013, 339, 1590–1593.

(8)

Eren, B.; Heine, C.; Bluhm, H.; Somorjai, G. A.; Salmeron, M. J. Am. Chem. Soc. 2015, 137, 11186-11190.

(9)

Poulston, S.; Parlett, P. M.; Stone, P.; Bowker, M. Surf. Interface Anal. 1996, 24, 811–820.

(10)

Schon, G. Surf. Sci. 1973, 35, 96–108.

(11)

Svintsitskiy, D. A.; Kardash, T. Y.; Stonkus, O. A.; Slavinskaya, E. M.; Stadnichenko, A. I.; Koscheev, S. V.; Chupakhin, A. P.; Boronin, A. I. J. Phys. Chem. C 2013, 117, 14588–14599.

(12)

Jernigan, G. G.; Somorjai, G. A. J.Catal. 1994, 147, 567-577.

(13)

Huang, T.; Tsai, D. Catal. Lett. 2003, 87, 4–9.

(14)

De Jong, K. P; Geus, J. W.; Joziasse, J. J. Catal. 1980, 441, 437–441.

(15)

Miro, E. E; Lombardo, E. A.; Petunchi, J. O. J. Catal. 1987, 104, 176–185.

(16)

Sun, B. Z.; Chen, W. K.; Xu, Y. J. J. Chem. Phys. 2010, 133, 154502-154508.

(17)

Habraken, F. H. P. M; Mesters, C. M. A. M; Bootsma, G. A. Surf. Sci. 1980, 97, 264–282.

32


Page 33 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44

ACS Catalysis

(18)

Domagala, M. E.; Campbell, C. T. Catal. Lett. 1991, 9, 65–70.

(19)

Szanyi, J.; Wayne, D. Catal. Lett. 1993, 21, 165–174.

(20)

Xu, F.; Mudiyanselage, K.; Baber, A. E.; Soldemo, M.; Weissenrieder, J.; White, M. G.; Stacchiola, D. J. J. Phys. Chem. C 2014, 118, 15902–15909.

(21)

Rojas, M. L.; Fierro, J. L. G; Tejuca, L. G.; Bell, A. T. J. Catal. 1990, 124, 41–51.

(22)

Biesinger, M. C.; Lau, L. W. M.; Gerson, A. R.; Smart, R. S. C. Appl. Surf. Sci. 2010, 257, 887– 898.

(23)

Espinós, J. P.; Morales, J.; Barranco, A.; Caballero, A.; Holgado, J. P.; González-Elipe, A. R. J. Phys. Chem. B 2002, 106, 6921–6929.

(24)

Goodby, B. E.; Pemberton, J. E. Appl. Spectrosc. 1988, 42, 754–760.

(25)

Figueiredo, R. T.; Martinez-Arias, A; Lopez Granados, M.; Fierro, J. L. J. Catal. 1998, 152, 146–152.

(26)

Agrell, J.; Boutonnet, M.; Melián-Cabrera, I.; Fierro, J. L. Appl. Catal. A Gen. 2003, 253, 201– 211.

(27)

Takagi-Kawai, M.; Soma, M.; Onishi, T.; Tamaru, K. Can. J. Chem. 1980, 58, 2132–2137.

(28)

Biesinger, M. C.; Brown, C.; Mycroft, J. R.; Davidson, R. D.; McIntyre, N. S. Surf. Interface Anal. 2004, 36, 1550–1563.

(29)

Biesinger, M. C.; Payne, B. P.; Lau, L. W. M.; Gerson, A.; Smart, R. S. C. Surf. Interface Anal. 2009, 41, 324–332.

(30)

Kobayashi, H.; Kubota, T.; Kawa, H.; Nakato, Y.; Nishiyama, M. Appl. Phys. Lett. 1998, 73, 933–935.

(31)

Chan, G. H.; Zhao, J.; Hicks, E. M.; Schatz, G. C.; Van Duyne, R. P. Nano Lett. 2007, 7, 1947– 1952.

(32)

Rice, K. P.; Paterson, A. S.; Stoykovich, M. P. Part. Part. Syst. Charact. 2015, 32, 373–380.

(33)

Larsson, E. M.; Langhammer, C.; Zorić, I.; Kasemo, B. Science 2009, 326, 1091–1094.

(34)

Tabib, P.; Adibi, Z.; Mazzotta, F.; Antosiewicz, T. J.; Skoglundh, M.; Gro, H.; Langhammer, C. ACS Catal. 2015, 5, 426-432.

(35)

Fredriksson, H. O. A.; Larsson Langhammer, E. M.; Niemantsverdriet, J. W. (Hans). J. Phys. Chem. C 2015, 119, 4085–4094.

(36)

Yin, Y.; Rioux, R. M.; Erdonmez, C. K.; Hughes, S.; Somorjai, G. A; Alivisatos, A. P. Science 2004, 304, 711–714.

(37)

Ghodselahi, T.; Zahrabi, H.; Saani, M. H.; Vesaghi, M. A. J. Phys. Chem. C 2011, 115, 2212622130.

(38)

Bingham, J. M.; Anker, J. N.; Kreno, L. E.; Van Duyne, R. P. J. Am. Chem. Soc. 2010, 132, 17358-17359.

(39)

White, B.; Yin, M.; Hall, A.; Le, D.; Stolbov, S.; Rahman, T.; Turro, N.; Brien, S. O. Nano Lett. 2006, 6, 2095-2098.

(40)

Smeets, P. J.; Groothaert, M. H.; van Teeffelen, R. M.; Leeman, H.; Hensen, E. J. M.; Schoonheydt, R. A. J. Catal. 2007, 245, 358–368.

(41)

Smeets, P. J.; Groothaert, M. H.; Schoonheydt, R. A. Catal. Today 2005, 110, 303–309.

(42)

Kondratenko, E. V.; Takahashi, N.; Nagata, N.; Ibe, M.; Hirata, H.; Takahashi, H. ChemCatChem 2015, 7, 3956–3962.

(43) van de Rotten, B. A.; Verduyn Lunel, S. M.; Bliek, A. Chem Eng. Sci. 2006, 61, 6981-6994. (44)

Chorkendorff, I.; Niemantsverdriet, J. W. Concepts of Modern Catalysis And Kinetics, Wiley-

33


ACS Catalysis

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

1

VCH: Weinheim, Germany, 2003.

2 3

4

34


Page 34 of 34

Cu Model Catalyst Dynamics and CO Oxidation Kinetics Studied by

Recommend Documents