Catalyst Architecture for Stable Single Atom Dispersion Enables Site-Specific Spectroscopic and Reactivity Measurements of CO Adsorbed to Pt Atoms, Oxidized Pt Clusters, and Metallic Pt Clusters on TiO2

Subscriber access provided by UNIVERSITY OF ADELAIDE LIBRARIES

Article

Catalyst architecture for stable single atom dispersion en-ables site-

specific spectroscopic and reactivity measurements of CO adsorbed

to Pt atoms, oxidized Pt clusters and metallic Pt clusters on TiO2

Leo DeRita, Sheng Dai, Kimberly Lopez-Zepeda, Nicholas Pham,

George W. Graham, Xiaoqing Pan, and Phillip Christopher

J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.7b07093 • Publication Date (Web): 13 Sep 2017

Downloaded from http://pubs.acs.org on September 14, 2017

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.

Journal of the American Chemical Society is published by the American Chemical

Society. 1155 Sixteenth Street N.W., Washington, DC 20036

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 17

Journal of the American Chemical Society

1

2

3

4

5

6

7

8

9

Catalyst architecture for stable single atom dispersion en-

ables site-specific spectroscopic and reactivity measure-

ments of CO adsorbed to Pt atoms, oxidized Pt clusters and

metallic Pt clusters on TiO₂

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

†

§

†

§

§,

∥

Leo DeRita , Sheng Dai , Kimberly Lopez-Zepeda , Nicholas Pham , George W. Graham , Xiaoqing Pan , and

Phillip Christopher^†,^⊥,

#,*

^†Department of Chemical and Environmental Engineering, ^⊥Program in Materials Science and Engineering, and ^#UCR Center for

Catalysis, University of California Riverside, Riverside, California 92521, United States

^§Department of Chemical Engineering and Materials Science and ^∥Department of Physics and Astronomy, University of California

Irvine, Irvine, California 92697, United States

Keywords: Single Atom Catalysts, Isolated Site Catalysts, Pt, CO oxidation, IR Spectroscopy

ABSTRACT: Oxide-supported precious metal nanoparticles are widely used industrial catalysts. Due to expense and rarity, developing syn-

thetic protocols that reduce precious metal nanoparticle size and stabilize dispersed species is essential. Supported atomically dispersed, single

precious metal atoms represents the most efficient metal utilization geometry, although debate regarding the catalytic activity of supported

single precious atom species has arisen from difficulty in synthesizing homogeneous and stable single atom dispersions and a lack of site-

specific characterization approaches. We propose a catalyst architecture and characterization approach to overcome these limitations, by

depositing ~1 precious metal atom per support particle and characterizing structures by correlating Scanning Transmission Electron Micros-

copy imaging and CO probe molecule infrared spectroscopy. This is demonstrated for Pt supported on anatase TiO₂. In these structures iso-

lated Pt atoms, Pt_iso, remain stable through various conditions and spectroscopic evidence suggests Pt_isospecies exist in homogeneous local

environments. Comparing Pt_isoto ~1 nm pre-oxidized (Pt_ox) and pre-reduced (Pt_metal) Pt clusters on TiO₂, we identify unique spectroscopic

signatures of CO bound to each site, and find CO adsorption energy is in the order: Pt_iso<<Pt_metal<Pt_ox. Pt_isospecies exhibited a 2-fold greater

turnover frequency for CO oxidation than 1 nm Pt_metalclusters, but share an identical reaction mechanism. We propose the active catalytic

sites are cationic interfacial Pt atoms bonded to TiO₂and that Pt_isoexhibits optimal reactivity because every atom is exposed for catalysis and

forms an interfacial site with TiO₂. This approach should be generally useful for studying the behavior of supported precious metal atoms.

specific activity and reduced catalyst costs.^10,11Additional interest in

isolated site Pt-group metal catalysts has been fueled by the idea

1. Introduction

Supported Pt-group metal catalysts are of critical importance due

that these materials could bridge the gap between heterogeneous

to their use in a broad range of chemical conversion technologies,^1,2

and homogeneous catalysis, potentially providing unique opportu-

with the highest demand in terms of metal mass coming from au-

nities to tune selectivity in difficult chemical conversions.^12–15

tomotive three-way catalytic convertors.^3,4Due to a naturally low

abundance of Pt-group metals in the earth’s crust,^5,6significant

efforts have been devoted to developing approaches that maximize

metal utilization efficiency thereby reducing costs and promoting

sustainability.⁷The primary approach to maximize metal utilization

efficiency is increasing the metal surface area to volume ratio by

decreasing the size of nanometer scale particles, leaving a greater

fraction of the total metal atoms in the catalyst available to drive

chemistry.⁸At the ultimate limit, this corresponds to single metal

atoms deposited on supports, a class of materials called single atom,

atomically dispersed, or isolated site catalysts, which exhibit perfect

metal utilization, where all atoms are exposed to reactants and

available for driving catalytic reactions.⁹The critical consideration

in motivating the development of isolated site Pt-group metal cata-

lysts is whether these materials will maintain, or enhance, the in-

herent reactivity of small nanoparticles to enable improvements in

Research into isolated Pt-group, and more broadly precious, metal

atoms on oxide supports has spanned decades with initial focus on

providing spectroscopic evidence of their existence. For example,

following the assignment of the Rh gem dicarbonyl species,

Rh(CO)₂, to CO bound to isolated Rh atoms on oxides using infra-

red (IR) spectroscopy, efforts were devoted to identifying the

structure, formation mechanism, and local geometry of these spe-

cies.¹⁶While initial research focused on the spectroscopic identifi-

cation of isolated precious metal atoms on oxide supports, the abil-

ity to directly observe single heteroatoms on supports by aberration

corrected scanning transmission electron microscopy (STEM) has

sparked renewed interest into the synthesis, characterization, and

potentially unique reactivity of isolated site precious metal cata-

lysts.¹⁷There is a growing body of literature reporting the reactivity

of isolated site precious metal catalysts for reactions including wa-

ter gas shift,^18–21CO₂hydrogenation,²²ethylene dimerization,²³

ACS Paragon Plus Environment

Journal of the American Chemical Society

Page 2 of 17

butadiene hydrogenation,¹³and NO reduction,²⁴suggesting a broad

cause these are also signatures of small supported Pt_oxclusters. Be-

1

2

3

4

5

6

7

8

and exciting range of potential applications. One system that has

received significant attention is the adsorption and oxidation of CO

on oxide supported isolated Pt atoms (Pt_iso). Attention stems from

the apparent simplicity of the reaction, but more relevantly from

the potential to use Pt_isospecies to reduce the massive demand of Pt

for use in automotive (mostly in diesel engines) CO oxidation

catalysis.^25,26Interestingly, significant variations exist in the report-

ed strength of interaction between CO and oxide supported Pt_iso,

and the reactivity of these species in CO oxidation. For example,

there are theoretical and experimental reports that CO binds weak-

er^8,27–31and stronger^32,33to Pt_isospecies compared to metallic Pt clus-

ter (Pt_metal) surfaces, with binding energies on Pt_isoranging from 20

to >200 kJ/mol depending on support and analysis approach. Simi-

larly, there are variations in reports regarding the comparative reac-

tivity of Pt_isoand Pt_metalclusters for CO oxidation, where Pt_isospecies

have been identified as the more and less reactive active site.^27,32–35

cause of these similarities, differentiation of Pt_isoand Pt_oxby XPS or

XAS requires maintained cationic charge and lack of Pt-Pt coordi-

nation in conditions where Pt_oxclusters reduce to Pt_metalclusters.

Furthermore, the existence of even a few Pt clusters in a sample can

lead to inconclusive results as the signal of all Pt species are aver-

aged in the calculation of coordination numbers in XAS.

Probe molecule infrared (IR) spectroscopy with CO is a widely

used characterization technique that allows probing of the local

structure, oxidation state, and coordination environment of sup-

ported precious metals.⁴²CO probe molecule IR is a sample-

averaged technique that also is site-specific due to the varying vi-

brational frequency of CO when adsorbed to different types of

supported metal sites, and it can be operated in a temperature pro-

grammed manner to extract information about the chemical reac-

tivity of distinct precious metal sites.²²Similar to issues with XAS or

XPS, distinguishing vibrational bands of CO when adsorbed to Pt_iso

and Pt_oxclusters requires consideration, as previous reports on

these species identify their stretching frequencies both existing in

the range ~2080-2130 cm^-1.^40,43–45The similar vibrational frequency

of CO when adsorbed to Pt_isoand Pt_ox, which is blueshifted (higher

frequency) from the stretching frequency of CO on Pt_metal(2030-

2100 cm^-1), arises from the similar cationic charge of Pt in both

structures.

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

Variations in the strength of interaction and reactivity of CO on

oxide supported Pt_isospecies are certainly related to the specific

metal oxide support composition, which acts as a ligand modifying

Pt_isospecies reactivity. However, given the range of conclusions

with inconsistency across similar classes of supports (i.e. reducible,

such as CeO₂, FeO_x, TiO₂) it is likely that variations in reported

activity and strength of interaction between CO and Pt_isoare also

related to challenges associated with the synthesis of stable Pt_iso

species on oxide supports and the unambiguous identification of

these species. While different approaches have been used to deposit

Pt_isospecies on oxide supports, it has been consistently reported

that agglomeration into small Pt clusters occurs during high tem-

perature oxidation (calcination),³⁵reduction,³⁶or exposure to reac-

tion conditions.³⁷The lack of Pt_isostability in response to standard

treatments used in the activation of Pt catalysts makes the rigorous

identification of Pt_isoreactivity difficult. For example, a lack of sta-

bility in high temperature calcination conditions necessitates mild

oxidation treatments, which results in retained existence of ligands

or counter ions on the catalyst that could influence reactivity and

mobility of the Pt complex.^31,35,38,39Furthermore, if Pt_isospecies can-

not be exposed to H₂at elevated temperature, there exists the pos-

sibility that small, oxidized Pt (Pt_ox) clusters co-exist on the sup-

port,⁴⁰which are difficult to differentiate from Pt_isospecies given the

similar cationic charge and local bonding environment of Pt in

these structures. Finally, a lack of stability under reaction condi-

tions makes it difficult to separate the reactivity of Pt_isospecies from

Pt_metalclusters. For rigorous analyses of the reactivity of supported

Pt_isospecies it is critical to develop strategies to maintain their struc-

ture under standard pretreatments used for supported Pt catalysts.

It is consistently seen across all characterization approaches that

the development of synthetic approaches to create stable Pt_isospe-

cies on oxide supports that survive through calcination, reduction,

and reaction conditions is required to unambiguously distinguish

Pt_iso, Pt_oxclusters and Pt_metalclusters. The importance of distinguish-

ing these species is substantiated by reports showing that Pt_oxspe-

cies ranging from small clusters containing a few Pt atoms to ex-

tended single crystal surfaces exhibit strong interactions with CO

and minimal CO oxidation reactivity,^46–48potentially obfuscating

measurements of Pt_isoreactivity if these species are not differentiat-

ed and distinctly compared.⁴⁹

Here we introduce a general approach to synthesize stable Pt_iso

catalysts by using ~5 nanometer (nm) diameter oxide nanoparti-

cles (TiO₂in this case) as supports and controlling Pt weight load-

ing and synthesis conditions to deposit ~1 Pt atom per support

particle. Using a correlated STEM and CO probe molecule IR

characterization approach it was shown that Pt remains site isolated

on TiO₂particles through high temperature (450 °C) calcination,

reduction in H₂(300 °C), exposure to CO oxidation reaction con-

ditions for ~3 days, and when aged on the shelf for ~1 month. CO

adsorption on Pt_isois characterized by a sharp (6-8 cm^-1full with at

half maximum, FWHM) IR band centered at 2112 cm^-1that does

not change frequency, width, or symmetry with coverage, while CO

adsorption on 1 nm diameter Pt_metal(2040-2100 cm^-1) and Pt_ox

(2100-2125 cm^-1) clusters is characterized by broader bands that

shift in frequency and shape with coverage. Temperature pro-

grammed analysis showed that the interaction strength between

CO and Pt species is in the order Pt_iso<< Pt_metal< Pt_ox. The reactivi-

ty of Pt_isoand Pt_metalwas compared in steady-state CO oxidation

measurements void of heat and mass transfer effects, which showed

that Pt_isoexhibited a 2-fold higher turnover frequency (TOF) com-

pared to 1 nm Pt_metalclusters at 200 °C, but an identical reaction

mechanism. We interpret the results in terms of the active site for

CO oxidation on TiO₂supported Pt being interfacial cationic Pt

atoms, demonstrating that when supported on TiO₂, Pt_isospecies

exhibit optimal reactivity on a per mass and site basis because every

In addition to the synthesis of stable Pt_isospecies, rigorous sample-

averaged and site-specific characterization approaches are needed

to differentiate Pt_iso, Pt_metalclusters and Pt_oxclusters. While STEM

imaging provides visual evidence of the presence of single atoms,

the statistically limited nature of the technique coupled with the

inability to provide 3-D representations of a support surface at

atomic resolutions from the 2-D projections that makeup images

suggest that STEM should be coupled with complimentary charac-

terization techniques. Photon-based spectroscopies, such as X-ray

absorption (XAS) and X-ray photoelectron spectroscopy (XPS),

have been widely used to identify the charge state and local coordi-

nation of supported metal catalysts.⁴¹The utility of these tech-

niques to differentiate characteristics of Pt_isospecies, such as a cati-

onic charge or coordination only to oxygen atoms is hindered be-

ACS Paragon Plus Environment

Page 3 of 17

Journal of the American Chemical Society

atom is exposed for catalysis and creates an interfacial site with the The high temperature and extended duration of the calcination

1

2

3

4

support. It is expected that the approach demonstrated here would

be generally useful for making rigorous assignments of spectroscop-

ic and reactivity behavior of isolated precious metals on various

oxides.

ensures removal of all remaining amine or nitrate ligands from the

TAPN precursor which are completely decomposed under these

conditions.⁵¹

2.3 STEM Imaging

5

STEM imaging was performed on a JEOL JEM-ARM300F Grand

ARM transmission electron microscope equipped with two spheri-

cal aberration correctors and a 300 kV cold field emission gun.

High-angle annular dark field (HAADF)-STEM images were rec-

orded using a convergence semi angle of 22 mrad and inner and

outer collection angles of 83 and 165 mrad, respectively. 20-30

images were collected of each sample to observe at least 100 Pt

species and enable convincing structural assignments of the CO IR

spectra. STEM imaging conditions were set to minimize the elec-

tron beam effect. A relatively small probe current 20 pA was used

for imaging. Meanwhile, the magnification was always kept below

8,000,000 and the acquisition time was less than 16 seconds. Single

Pt atoms were observed to be stable under these conditions.

6

7

8

9

2. Experimental Methods

2.1 Materials

High purity (99.995%) Tetraammineplatinum(II) nitrate (TAPN)

was purchased from Sigma-Aldrich (#482293) and used as the Pt

precursor. 5 nanometer diameter anatase TiO₂(99.5%) crystals

with high surface area (290 m²/g) were purchased from US Re-

search Nanomaterials (Stock # US3838) and used as the support in

this study. Reagent grade NH₄OH (28-30% concentration) used in

catalyst synthesis was obtained from Acros Organics

(#423300250). SiO₂gel (Davisil Grade 62, pore size 150 Å, 60-200

mesh, Part #243981, Sigma Aldrich) and acid purified SiO₂(Sigma

Aldrich, #84880) were used for catalyst co-impregnation and cata-

lyst dilution, respectively. For the IR and steady-state reactivity

experiments, the following gases were used: 10% H₂/Ar ( 2%,

Airgas), He (UHP, 99.999%, Airgas), 10% O₂/He ( 2%, Airgas),

and 10% CO/He ( 2%, Airgas) in an Al tank to prevent metal

carbonyls from entering reactant streams.

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

2.4 IR Characterization

Prior to collecting CO probe molecule IR spectra, catalysts were

loaded into a Harrick High Temperature Reaction chamber (ZnSe

windows) mounted inside a ThermoScientific Praying Mantis dif-

fuse reflectance adapter set in a Nicolet iS10 FTIR spectrometer

with a Mercury Cadmium Telluride (MCT) detector cooled by

liquid nitrogen. Mass flow controllers (Teledyne Hastings) were

used to control the gas flow rates across the reactor bed and all

gases were first passed across both an isopropyl alcohol-liquid ni-

trogen cold trap held at -80 °C and a glass trap filled with Drierite

desiccant to remove trace moisture. The surface temperature in the

catalyst bed was calibrated using an optical pyrometer due to the

known gradient in the bed of the Harrick reactor.^52,53

2.2 Catalyst Synthesis

The synthesis of catalysts containing a predominance of Pt_isospe-

cies was achieved by a wet impregnation technique called strong

electrostatic adsorption (SEA)⁵⁰using low Pt weight loadings

(0.025-0.15 wt%), large synthesis volumes to promote homogene-

ous deposition of Pt on the support, and controlled solution pH (8-

12.5). In a typical synthesis of 0.025 wt% catalyst containing exclu-

sively Pt_isospecies, 1 g of TiO₂was crushed, sieved to a powder finer

than 75 µm, and dried in a vacuum oven at 120 °C overnight prior

to synthesis. 25 mL of deionized water was mixed with 75 mL of

NH₄OH to dissolve the TiO₂support, resulting in a solution pH of

12.2. Separately, 10 mg of TAPN was dissolved in 5 mL of deion-

ized water from which 300 μL were taken and added to 25 mL of

NH₄OH to create a precursor solution with a pH of 12.2. The 25

mL TAPN solution was injected over 12.5 hours into the support

solution while constantly stirring to achieve a final total solution

volume of 125 mL with a surface loading (see equation 1) of 2900

m²/L. Following the slow precursor addition, the final solution was

heated to 70 °C until completely dried.

CO probe molecule IR spectroscopy was used to identify unique

vibrational signatures of CO adsorbed to various Pt structures. To

examine the spectral features of CO on each structure, as well as the

stability of CO and reactivity in an O₂flow, the following protocol

was used. First, an in-situ oxidation (300 °C for 30 min in a 10%

O₂/He mixture at 50 sccm) or reduction (240 °C for 1 hour in pure

H₂, 50 sccm) pre-treatment was performed to allow differentiation

of Pt_oxclusters from Pt_metalclusters and Pt_isospecies. Next, catalysts

were cooled to room temperature in He flowing at 100 sccm. Once

at room temperature, CO was adsorbed until saturation coverage

by flowing 10% CO/He at 50 sccm for 10 minutes, where it was

observed that bands associated with CO adsorption on Pt no long-

er changed. The system was flushed for 2 minutes in He at 100

sccm to remove any gas phase CO, which has vibrational bands that

overlap with bands associated with CO adsorbed on cationic Pt

species. Next, the stability of remaining chemisorbed CO was

probed either in a temperature programmed desorption (TPD)

experiment or in a temperature programmed oxidation (TPO)

experiment. In the TPD experiment, the temperature was ramped

from room temperature to 500 °C at a rate of ~7 °C/min in 100

sccm of He flow. Alternatively in the TPO experiment, oxygen was

introduced (50 sccm of 10% O₂/He) to

ꢀꢁꢂꢀ_!"##$%&∗ ꢀꢁꢂꢃꢄꢅꢆ ꢀꢁꢂꢀ_!"##$%&

1) ꢀꢁꢂꢃꢄꢅꢆ ꢀꢁꢂꢃꢄꢅꢆ =

ꢀꢁꢂꢃꢄꢀ_!"#$%&"'

For the synthesis of 1 wt% Pt catalysts containing Pt clusters and no

Pt_isospecies, an incipient wetness (dry) impregnation (DI) proce-

dure was used. In a typical synthesis of 1 g of catalyst, 20 mg of

TAPN was dissolved in 1200 μL of deionized water and the solu-

tion was added to 990 mg of TiO₂(dried and sieved as mentioned

above), forming a paste. The paste was heated to 70 °C until com-

pletely dried. The dried catalysts were placed in a vacuum oven at

120 °C overnight and sieved to 75 µm.

All catalysts were calcined in a tube furnace at 450 °C for 4 hours in

flowing air after ramping up to temperature at a rate of 10 °C/min.

ACS Paragon Plus Environment

Journal of the American Chemical Society

Page 4 of 17

(A)

(B)

1

2

3

4

5

6

7

8

Dry Impregnation

Strong Electrostatic Adsorption

(NO₃)^-

OH

O^-

[(NH₃)₄Pt]⁺²

0.05 wt%

Pt/TiO₂

0.05 wt%

Pt/TiO₂

1 wt% Pt/TiO₂

O^-

(NO₃)^-

OH

(NO₃)^-

TiO₂Surface Area = 290 m²/g

9

5 nm

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

pH

More Basic

More Neutral

Figure 1. Synthesis Approaches for producing site-isolated Pt and Pt clusters on TiO₂. (A). Schematic of the strong electrostatic adsorption

based synthetic protocol used to deposit 1 Pt atom per 5 nm diameter TiO₂particle through controlling weight loading, solution volume, and

solution pH. By tuning the solution pH, a Coulombic attraction between the surface oxygen anion (O^-) and the cationic precursor

[(NH₃)₄Pt]⁺²promotes the deposition of 1 Pt atom per TiO₂particle. (B). A schematic of the dry impregnation approach for producing ex-

clusively Pt clusters, where higher weight loadings, no control of pH and smaller solution volumes are used to deposit the Pt precursor onto

TiO₂particles.

the chemisorbed CO still held at room temperature and a tempera-

ture ramp to 500 °C at a ramp rate of 10 °C/min followed. Spectra

were recorded throughout the entirety of experiments in Kubelka

Munk (KM) units with 32 scans and 0.482 cm^-1data spacing being

averaged into a spectrum, allowing half-minute increments be-

tween each measurement.

increment steps over a period of 2 minutes, followed by holding at a

given temperature for 2 hours to ensure steady state signals were

reached in the mass spectrometer (MS). Partial pressure depend-

ent measurements were made utilizing a home-built programmable

Arduino controller that automated a solenoid valve control that

allowed the reactants to bypass the reactor for baseline measure-

ments in the MS before and after every change in pressure. An iden-

tical pretreatment to the activation barrier measurements preceded

the partial pressure dependent measurements including the 10

hours on stream at 200 °C. Following the activation period, instead

of cooling, the partial pressure dependent measurements were exe-

cuted at 200 °C. The total flowrate for all partial pressure measure-

ments was 200 sccm in a balance of He. The CO partial pressure

dependence was run first, beginning with the highest CO feed ratio

of 1.5% CO to 0.5% O₂and ending with the lowest CO feed ratio of

0.25% CO to 0.5% O₂. At the start and end of each feed ratio, the

reactants were flowed for 20 min bypassing the reactor to get a

baseline signal in the MS. In between baseline measurements, the

reactants were flowed across the catalyst bed for a period of 2 hours

to obtain a steady state signal. Immediately following the lowest

CO feed composition, while still at 200 °C, the O₂partial pressure

dependent measurements began with the highest O₂feed ratio

(0.5% CO to 1.5% O₂) following a 40 min bypass period. The al-

ternating bypass/reactor cycle continued for 7 steady state meas-

urements with sequentially decreasing O₂feed composition and

constant CO pressure, ending with a feed ratio of 0.5% CO to

0.25% O₂. For both sets of reaction experiments and for all cata-

lysts, the reported kinetic measurements are an average of at least 3

unique measurements (i.e. new catalyst preparations were loaded

into the reactor) to ensure repeatability and consistency. Tempera-

ture was controlled in the home-built furnaces using programmable

PID controllers (Omega CN7800). Flowrates were controlled by

calibrated mass flow controllers (Teledyne Hastings). Gas streams

were cleaned of humidity prior to the reactor using a Drierite bed

(tests with the addition of an isopropyl alcohol-liquid nitrogen trap

showed no effect on rate), and Al-lined CO tanks were purchased

to prevent metal carbonyl formation and ensure clean CO streams.

All effluent gases were measured using calibrated online mass spec-

trometry (Halo 201, Hiden Analytical).

2.5 Steady State CO Oxidation Kinetics

Differential reactor measurements were made in a glass tubular

packed-bed reactor to directly measure and compare the steady

state CO oxidation reactivity of the various prepared catalysts. In all

cases, a half-inch glass tube was used with a 4:1 length to diameter

ratio of the catalyst bed to minimize the possibility of channeling.

Calcined Pt/TiO₂catalysts were sieved to < 75 µm and co-

impregnated with inert SiO₂gel at a 1:4 ratio of catalyst to SiO₂.

The mixture was sonicated for 30 minutes in a round bottom flask

with water and stirred overnight. The flask was then transferred to a

rotary evaporator and vacuum dried over a 30 minute period at 110

rpm. The entire batch of co-impregnated catalyst was then further

diluted 5:1 by an acid purified SiO₂(Sigma Aldrich, #84880) and

loaded into the packed bed reactor. The dilution and co-

impregnation steps were used to minimize (1) the pressure drop

that would occur with gas flow through a packed bed of 5 nm diam-

eter TiO₂particles, (2) the size of domains containing 5 nm TiO₂

particles, which would introduce mass transfer and diffusion effects

on the kinetic rates, and (3) hot-spot formation that is associated

with this highly exothermic reaction. No pressure drop across the

catalyst bed was observed following the co-impregnation of the

catalyst and across all flow rates of interest the reaction rate stayed

constant (Figure S1) suggesting no mass transfer limitations (i.e.

kinetic measurements were made in a reaction limited regime).

For steady-state CO oxidation measurements, 50 mg of catalysts

were loaded into the reactor and pre-treated with a 2 hour reduc-

tion in pure H₂at 20 sccm and 300 °C. The catalyst was cooled to

200 °C in 160 sccm He and then exposed to reaction conditions

(1% CO, 1% O₂, 98% He at 200 sccm total flowrate) for 10 hours

to allow steady state to be reached; during the 10 hour period reac-

tion rates varied by less than 5%. Following the activation period,

the reactor was cooled to 140 °C in the same composition and acti-

vation barrier measurements were made using 10 °C temperature

ACS Paragon Plus Environment

Page 5 of 17

Journal of the American Chemical Society

ously mixed solutions and controlled TiO₂-Pt cation interactions, it

3. Results

1

2

3

was expected that there was a high probability to produce stable

Pt_isospecies.

3.1 Synthesis and vibrational band assignments

of CO adsorbed to Pt_iso

As a comparison for the designed synthesis of Pt_isocatalysts, a 1 wt%

Pt/TiO₂catalyst was prepared by incipient wetness (dry) impreg-

nation (DI) with the goal of exclusively producing Pt clusters, Fig-

ure 1(B). The 1 wt% loading dictates that on average ~8.4 Pt atoms

would be deposited per TiO₂particle, with low solution volume

and lack of pH modification eliminating the expected driving force

for Pt_isoformation. Even with the lack of specific efforts to control

Pt particle size in this synthesis, nanometer and sub-nanometer

diameter particles were expected to form due to the required migra-

tion between TiO₂particles for particle sintering and the relatively

small number of Pt atoms deposited per TiO₂particle.

4

5

6

7

8

9

pH 8

2,900 m²/L

3,600 m²/L

14,500 m²/L

(A)

(B)

pH 10

pH 11

pH 12.5

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

Due to its high sensitivity to local bonding environment, CO probe

molecule IR spectroscopy was utilized to analyze how synthesis

parameters influenced the resulting Pt structure. Characterization

of structural properties and charge of supported Pt catalysts using

CO as a probe molecule with IR is abundant in literature.^42,57For

supported Pt_metalparticles there are well-established values for vibra-

tional frequency of adsorbed CO.⁵⁸When CO is adsorbed to Pt_metal

sites the vibrational frequency increases from 2040 to 2100 cm^-1as

Pt-Pt coordination number increases from ~5 (characteristic of

corners or defect sites on Pt particles) to 9 (characteristic of ex-

tended (111) surfaces on Pt particles).⁵⁸In contrast, when CO is

adsorbed to oxide supported cationic Pt species (Pt_oxclusters, Pt_iso

species, or Pt coordinated with oxidizing ligands) the CO vibra-

tional frequency is blueshifted from the Pt_metaladsorption sites,

typically to >2100 cm^-1.⁵⁹The blueshifted frequency for CO ad-

sorbed to cationic Pt sites compared to Pt_metalsites derives from the

decreased charge transfer to CO and CO bond distance being clos-

er to the gas phase CO distance.⁶⁰We hypothesized that for cata-

lysts synthesized using the SEA approach described above, which

were pretreated through harsh calcination (4 hours at 450 °C in air)

and in-situ reduction (1 hour at 240 °C in H₂), any remaining IR

bands of CO adsorbed to Pt sites with stretching frequencies >

2100 cm^-1must be due to the presence of Pt_isospecies, as all oxidiz-

ing ligands from synthesis would be removed by the calcination and

any Pt_oxclusters would be reduced to Pt_metalclusters during the re-

duction.

2200

2100

2000

1900

2200

2100

2000

1900

Wavenumber (cm^-1)

Figure 2. Surface loading and pH optimization for producing

Pt_iso. (A). IR spectra of CO adsorbed at saturation coverage and

room temperature to 0.05 wt% Pt/TiO₂SEA catalysts prepared at a

pH of 12 and precursor solution addition rate of 2 mL/hr and varying

surface loadings. (B). IR spectra of CO adsorbed at saturation cover-

age and room temperature to 0.05 wt% Pt/TiO₂SEA catalysts pre-

pared a 3,600 m²/L surface loading, precursor solution addition rate

of 2 mL/hr, and varying solution pH’s. All catalysts were reduced at

240 °C in H₂prior to CO adsorption. The spectra were normalized to

the intensity of the band at 2112 cm^-1, assigned to CO adsorbed to

Pt_iso, to enable direct comparison of the influence of synthetic condi-

tions on the relative proportion of CO adsorption sites. Catalysts

were pretreated in flowing H₂for 1 hour at 240 °C.

It was hypothesized that the preparation of exclusively Pt_isospecies

with high stability could be achieved by using small oxide nanopar-

ticles as supports and depositing ~1 Pt atom per support particle. In

this geometry, the required migration between support particles

should minimize agglomeration to form Pt multimers, even if Pt

atoms were mobile on the oxide surface. To synthesize this struc-

ture we used ~5 nanometer diameter anatase TiO₂particles (sur-

face area 290 m²/g), and low Pt weight loadings that correspond to

less than 1 Pt atom per TiO₂particle (0.025-0.05 wt% correspond-

ing to 0.2-0.4 Pt atoms/TiO₂particle). If a stochastic distribution

of Pt atoms were deposited onto TiO₂particles, there would be a

significant fraction of TiO₂particles containing >1 Pt atom, see

Figure 1(A). To mitigate the clustering of Pt atoms that would be

expected on a stochastic basis and promote a homogeneous deposi-

tion of the Pt atoms across the support particles, large synthesis

volumes were coupled with Pt precursor addition over long time

periods, creating an evenly mixed and dilute slurry. Principles from

the strong electrostatic adsorption (SEA) wet impregnation tech-

nique were used to promote repulsive interactions between Pt ions

in solution and attractive interactions between Pt ions and the TiO₂

surface.^50,54–56By modifying the TiO₂support surface through pH

adjustment by the addition of NH₄OH, hydroxyls were deproto-

nated to form O- anions at the TiO₂surface that Coulombically

attract [(NH₃)₄Pt]⁺², as shown in Figure 1(A). Through a combi-

nation of small TiO₂particles, low Pt weight loadings, homogene-

Figure 2(A) shows the IR spectra of CO adsorbed at saturation

coverage and room temperature to 0.05 wt% SEA catalysts synthe-

sized at a pH of 12, Pt precursor addition rate of 0.025 mg of Pt per

hour, and varying surface loading (inversely proportional to syn-

thesis solution volume, or qualitatively dilution). The spectra were

normalized to the intensity of the CO stretching band at 2112 cm^-1

to compare the relative amount of CO adsorbed to Pt_metalsites

(2040-2090 cm^-1). It can be seen that as the surface loading was

decreased (more dilute synthesis condition), the intensity of the

band associated with CO adsorption at a cationic Pt site (2112 cm^-

¹) increases relative to the intensity of the broad band associated

with CO adsorption on Pt_metalsites.

Figure 2(B) shows the spectra of CO adsorbed at saturation cover-

age to 0.05 wt% SEA catalysts synthesized at a Pt precursor addi-

tion rate of 0.025 mg of Pt per hour, surface loading of 3,600 m²/L

and varying pH between 8 and12.5.

ACS Paragon Plus Environment

Journal of the American Chemical Society

Page 6 of 17

8.00

0.10

0.08

0.06

0.04

0.02

0.00

4.00

3.20

2.40

1.60

0.80

0.00

1.0 wt%

0.05 wt%

0.15 wt%

(C)

1

2

(A)

(B)

CO on Pt_iso

6.40

4.80

3.20

1.60

0.00

1.25 Pt

atoms/ TiO₂

Particle

8.37 Pt

atoms/ TiO₂

Particle

0.42 Pt

atoms/ TiO₂

Particle

3

4

5

6

7

8

9

CO on Pt_metal

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

2200

2100

2000

1900

2200

2100

2000

1900

2200

2100

2000

1900

Wavenumber (cm^-1)

(E)

(F)

(D)

Figure 3. Correlative IR and STEM for CO vibrational band assignments on Pt_isoand Pt_metal. (A-C) IR spectra of CO adsorbed at saturation

coverage and room temperature to 0.05 wt% Pt/TiO₂SEA, 0.15 wt% Pt/TiO₂SEA and 1 wt% Pt/TiO₂DI catalysts, respectively, that had been re-

duced at 240 °C in H₂prior to CO adsorption. (D-F) Representative STEM images of the same 3 catalysts after reduction ex-situ, where IR spectra in

(A-C) correlate to STEM images of the same catalyst in (D-F). In the STEM images, Pt_isospecies with diameter of ~175 pm are circled in yellow,

while clusters with diameter ~ 1 nm are circled in red.

It was observed that as the synthesis pH was increased the catalysts

trended toward containing a single CO adsorption site with a

symmetric CO stretching band at a frequency of 2112 cm^-1and a

FWHM of < 10 cm^-1, with optimized synthetic conditions consist-

ently producing FWHM of 6-8 cm^-1. Catalysts synthesized at closer

to neutral pH contained Pt_metalCO adsorption sites and broadened

asymmetric bands centered at 2112 cm^-1, as seen in Figure 2B.

6-10 cm^-1was observed. Given that in all cases the catalysts were

pre-treated by H₂under conditions known to reduce Pt_oxclusters

into Pt_metalclusters, the retained existence of a CO adsorption site

with a stretching frequency characteristic of adsorption on cationic

Pt atoms, and the sharpness of the CO band, suggest an assignment

of the 2112 cm^-1vibrational mode to CO adsorbed to Pt_isospecies

on the anatase TiO₂support.

A series of Pt/TiO₂catalysts were prepared with the optimized

conditions of Pt precursor addition rate of 0.025 mg of Pt per hour,

surface loading of 3,600 m²/L, pH of 12.2 and varying Pt loading

from 0.025 – 0.5 wt%. The IR spectra of CO adsorbed at saturation

coverage to this series of catalysts are shown in Figure S2. It can be

seen that when there was nominally less than 1 Pt atom/TiO₂par-

ticle during the synthesis (0.025-0.05 wt%) a single, sharp CO band

with a frequency at 2112 cm^-1was observed. When the Pt weight

loading was increased above 1 Pt atom/TiO₂particle (> 0.1 wt%) a

second broad CO band formed with a frequency of 2040-2090 cm^-

¹, signifying CO adsorbed to Pt_metalsites. The ratio of the band in-

tensity at 2040-2090 cm^-1increased relative to the 2112 cm^-1band

as the Pt weight loading was increased.

To support the IR assignment of CO adsorbed to Pt_isoon anatase

TiO₂, a correlated IR-STEM analysis was executed. We focused on

3 samples: (1) 0.05 wt% catalyst synthesized by the optimized SEA

protocol, (2) 0.15 wt% catalyst synthesized by the identical SEA

protocol, and (3) 1 wt% catalyst synthetized by the DI approach.

Both SEA catalysts were prepared using pH of 12.2, surface loading

of 2900 m²/L and Pt precursor addition rate of 0.025 mg of Pt per

hour. For these samples we collected CO probe molecule IR spec-

tra at saturation coverage, see Figures 3(A-C), and 20-30 STEM

images of the same materials, examples of which are shown in Fig-

ures 3(D-F), after they had been calcined and reduced through

identical protocols prior to both characterizations. In ~30 STEM

images of the 0.05 wt% SEA catalyst, Pt_isospecies were the predom-

inantly observed Pt species with only ~3 Pt clusters with diameter

<1 nm observed, Figure 3(D). An average diameter of 0.176 nm

was observed for the Pt_isospecies, see Figure S3. The measured

diameter of the Pt_isospecies cannot be directly related to the Pt

atomic radius, but instead shows that the measured size of Pt_iso

To summarize the influence of synthetic protocol on the CO probe

molecule IR spectra, it was found that when Pt was deposited onto

5 nm anatase TiO₂particles under highly dilute conditions, at high

pH, and with weight loadings lower than 1 Pt atom/TiO₂particle, a

single CO stretching band with frequency 2112 cm^-1and FWHM of

ACS Paragon Plus Environment

Page 7 of 17

Journal of the American Chemical Society

species is larger than the resolution of the instrument and thus, Pt_isoO₂), cooled to room temperature, and exposed to CO again for

1

2

3

4

5

6

7

8

species can be distinguished from dimers or trimers of Pt. The

STEM images identifying Pt_isospecies are in excellent agreement

with the IR spectra of the same CO saturated sample, Figure 3(A),

where the intensity of the band at 2112 cm^-1was ~180-times more

intense than the broad band observed between 2040-2090 cm^-1.

For the 0.15 wt% SEA catalyst, STEM imaging showed a significant

increase in the number of Pt clusters (0.84 nm average diameter)

compared to the 0.05 wt% SEA catalyst and the maintained exist-

ence of Pt_isospecies, Figure 3(E), in good agreement with the IR

spectra of the CO saturated catalyst, Figure 3(B), which showed an

increased relative intensity of the CO stretching band at 2040-2090

cm^-1compared to the 0.05 wt% SEA catalyst. Finally, STEM images

of the 1 wt% DI catalyst showed the existence of predominantly

clusters with an average diameter of 1.1 nm and a few larger parti-

cles with an average diameter of 4.3 nm, Figure 3(F), corroborating

the IR of the CO saturated catalyst in Figure 3(C), which showed

almost exclusively CO stretching band intensity between 2040-

2090 cm^-1.

characterization by IR. By performing the experiments sequentially

on the same catalysts, spectra of CO adsorbed to Pt_metaland Pt_ox

clusters could be compared to CO adsorbed to Pt_isospecies that

have been pre-oxidized or pre-reduced.

In Figure 4(A) it is seen that following reduction, the spectrum of

CO adsorbed at saturation coverage to the 1 wt% DI catalyst shows

two predominant CO stretching bands overlapping in the region

associated with CO adsorbed to Pt_metalsites centered at 2077 cm^-1,

assigned to well-coordinated (WC) Pt sites, and 2058 cm^-1, as-

signed to under-coordinated (UC) Pt sites.^61,62The spectrum

showed almost no intensity in the frequency range associated with

CO adsorbed to cationic Pt sites (>2100 cm^-1), suggesting fully

reduced Pt clusters.

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

Following oxidation of the 1 wt% DI catalyst, the spectrum of ad-

sorbed CO showed a decrease in the intensity of bands associated

with CO bound to Pt_metalsites (2040-2090 cm^-1) as compared to the

reduced catalyst, and the appearance of a broad band associated

with CO adsorbed to cationic Pt sites (>2100 cm^-1), see Figure

4(A). Peak deconvolution using a least-squares fit of multiple

Gaussian peaks showed that the CO stretching band associated

with adsorption on cationic Pt sites was centered at ~2118 cm^-1and

had a FWHM of 25-30 cm^-1, Figure S4. STEM imaging of the 1%

DI catalyst following the reductive and oxidative treatment showed

essentially identical Pt structure geometries, containing predomi-

nantly Pt clusters with an average of ~1 nm diameter and a few

larger Pt particles with an average ~4 nm diameter, Figure S5. Giv-

en the lack of structural change in Pt particles following oxidative

and reductive treatment, the broad CO stretching band between

~2090-2140 cm^-1in Figure 4A is assigned to CO adsorption on Pt_ox

clusters, consistent with previous reports.^45,63The CO stretching

bands at frequencies below 2090 cm^-1observed in Figure 4(A)

following oxidative treatment are assigned to Pt structures (or por-

tions of Pt structures) that were reduced by the flowing CO, as

previously observed for Pt/TiO₂catalysts.⁶⁴However, from the

combined IR and STEM data, which showed no significant struc-

tural changes of Pt following oxidative and reductive treatment,³⁶it

is clear that CO only partially reduced the Pt_oxclusters. This is an

important point that will be addressed in the discussion section, as

the only partial reduction of oxidized Pt clusters by CO suggests

the possible existence of Pt structures containing reduced Pt metal

species in addition to retained oxygen and is supported by the ob-

servation of CO stretching bands at frequencies associated with

both adsorption on metallic and cationic Pt atoms.

STEM imaging showed conclusive evidence of the retained exist-

ence of predominantly Pt_isospecies on the 0.05 wt% SEA catalyst

following calcination and reduction, as hypothesized in the design

of the synthesis approach. Furthermore, the correlated STEM-IR

analysis provides convincing evidence that the sharp CO stretching

band at 2112 cm^-1can be assigned to CO adsorbed on Pt_isoon ana-

tase TiO₂.

3.2 Site-specific signatures of CO adsorbed to

Pt_iso, Pt_metaland Pt_ox

8.0

6.0

4.0

2.0

0.0

0.0008

0.0006

0.0004

0.0002

0

0.08

0.06

0.04

0.02

0.00

Pre-Oxidized

Pre-Reduced

Pre-Oxidized

Pre-Reduced

(A)

(B)

CO on Pt_iso

CO on Pt_metal

CO on

Pt_ox

x100

2200

2100

2000

1900

2200

2100

2000

1900

Wavenumber (cm^-1)

Figure 4. Differentiating CO adsorbed to Pt_iso, Pt_metaland Pt_ox. (A)

CO adsorbed at room temperature and saturation coverage to a 1 wt%

Pt/TiO₂DI catalyst that had been pre-reduced (red), and pre-oxidized

(black), sequentially. (B) CO adsorbed at room temperature and

saturation coverage to a 0.05 wt% Pt/TiO₂SEA catalyst that had been

pre-reduced (red), and pre-oxidized (black), sequentially. The intensi-

ty of the spectrum collected from CO on the pre-oxidized catalyst was

multiplied by 100 to allow comparison.

In contrast to the behavior of Pt clusters seen in Figure 4(A), where

the CO stretching bands assigned to adsorption on cationic Pt sites

only appeared following oxidation pre-treatments, for the 0.05 wt%

SEA catalyst it was observed that the predominant CO stretching

band at 2112 cm^-1, which we assigned to CO adsorbed to Pt_iso, is

present following pre-treatment in a reducing atmosphere, see Fig-

ure 4(B). However, following pre-oxidation of the catalyst and CO

adsorption the overall intensity of the CO stretching bands in the

IR spectrum significantly decreased (100-fold), resulting in bands

at 2040-2090 cm^-1and 2090-2130 cm^-1, which appear similar to the

spectrum of pre-oxidized 1 wt% DI shown in Figure 4(A). It is pro-

posed that Pt_isospecies either exhibit preferential adsorption of O₂

compared to CO or that vacancies in the TiO₂lattice induce migra-

tion that poisons the Pt_isospecies.^34,65Thus, following the oxidative

treatment, Pt_isospecies are poisoned and do not allow CO to ad-

With the demonstrated stability of Pt_isospecies on 5 nm TiO₂parti-

cles following reductive treatment, we chose the 1 wt% DI (con-

taining predominantly Pt clusters) and 0.05 wt% SEA (containing

predominately Pt_isospecies) catalysts to identify spectroscopic sig-

natures that differentiate CO adsorbed to Pt_isoand Pt_oxclusters,

which both contain cationic Pt atoms. To achieve this, both cata-

lysts were first pre-treated for 1 hour at 240 °C in flowing H₂, then

exposed to CO at room temperature to saturation coverage and

characterized by IR. The catalysts were then oxidized to remove

adsorbed CO and oxidize the Pt structures (280 °C, 30 min in 10%

ACS Paragon Plus Environment

Journal of the American Chemical Society

Page 8 of 17

0.12

0.09

0.06

0.03

0.00

6.0

4.5

3.0

1.5

0.0

3.0

2 min He

10 min He

1 min O₂

2 min O₂

4 min O₂

10 min O₂

70°C

2 min He

4 min He

10 min He

20 min He

(B)

1%

(C)

2 min He

10 min He

1 min O₂

2 min O₂

4 min O₂

10 min O₂

70°C

(A)

0.05%

1%

1

2

3

4

5

6

7

8

He

O₂

2.3

1.5

0.8

0.0

He

O₂

120°C

230°C

300°C

160°C

230°C

350°C

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

TPO

2200

2100

2000

1900

2200

2100

2000

1900

2200

2100

2000

1900

Wavenumber (cm^-1)

Figure 5. Site-specific signatures of CO adsorbed to Pt_iso, Pt_metaland Pt_ox. (A). IR spectra of CO adsorbed at room temperature and saturation cov-

erage on a pre-reduced 0.05 wt% SEA catalyst during He flushing. (B). CO adsorbed to a reduced 1 wt% DI catalyst during a He and O₂flush at room

temperature and a TPO ramp. (C). CO adsorbed to an oxidized 1 wt% DI catalyst during the identical protocol as (B).

sorb, while the few Pt clusters in this catalyst exist in a combination

of reduced and cationic atoms, similar to the 1 wt% DI catalyst.

In contrast to the stability of CO exhibited on Pt clusters, Figure

5(A) shows IR spectra collected following saturation of Pt_isosites by

CO on a reduced 0.05 wt% SEA catalyst under flowing He at room

temperature. It was observed that all CO desorbed from Pt_isosites

within 30 minutes following first order unimolecular desorption

kinetics as shown in Figure S7, thus no temperature ramp was exe-

cuted. It is obvious that CO binds more weakly to the Pt_isosites

compared to Pt_metalor Pt_oxsites. In addition to the low stability of

CO observed on Pt_isosites, throughout desorption the CO stretch-

ing band remained symmetric and the FWHM stayed constant at

~8 cm^-1. Furthermore, as CO desorbed, the vibrational frequency

was constant. The absence of a coverage dependent shift in the

stretching frequency of CO on Pt_isosites shows there is no interac-

tion between adsorbed CO molecules and CO must therefore be

spatially isolated on atomically dispersed Pt atoms, in contrast to

the behavior observed on Pt_metaland Pt_oxsites in Figure S6.⁶⁷Lastly,

the retained symmetry and FHWM of the vibrational band of CO

on Pt_isoduring desorption is strong evidence that the Pt-CO species

are homogeneous in nature, likely all located at a common adsorp-

tion site on the TiO₂support.⁶⁸

To briefly summarize, through the use of the 1 wt% DI and 0.05

wt% SEA catalysts, which contain Pt clusters and Pt_isospecies, re-

spectively, and controlling oxidative or reductive pretreatments, we

can differentiate the IR spectra of CO adsorbed to Pt_iso, Pt_oxclus-

ters, and Pt_metalclusters. This is critical, as CO adsorption on both

cationic Pt species shows bands with similar stretching frequencies

(2112 cm^-1on Pt_isoand 2118 cm^-1on Pt_oxclusters), with the main

difference being the width of the band (<10 cm^-1on Pt_isoand 25-30

cm^-1on Pt_ox), making them otherwise difficult to differentiate. The

rigorous assignment of vibrational bands associated with CO ad-

sorption on Pt_iso, Pt_oxand Pt_metalsites enables comparison of the CO

adsorption strength and characteristics of the CO desorption pro-

cess on each adsorption site through TPD experiments in the IR.

Figure S6(A) and S6(B) show the IR spectra associated with vari-

ous temperatures recorded during TPD experiments (β = 7

°C/min) performed on 1 wt% DI catalysts that had been pre-

reduced and pre-oxidized, respectively. On the pre-reduced cata-

lysts, desorption of CO was observed to initiate from Pt_metalsites at

~50 °C and be mostly completed by 300 °C, Figure S6(A). The

desorption temperature was higher than traditionally seen in TPD

experiments of CO from Pt clusters. We attribute this discrepancy

to the time needed to average data collected in a given spectrum,

causing each spectrum to represent a range of temperatures.⁶²With

increasing temperature, a characteristic redshift in CO stretching

frequency occurred as the band intensity decreased due to the de-

creased dipole-dipole interactions of neighboring CO molecules on

the Pt_metalsurface. For the pre-oxidized 1 wt% DI catalyst, it was

observed that CO was bound more strongly to cationic Pt_oxsites

(>2100 cm^-1) as compared to Pt_metalsites (<2100 cm^-1), with signifi-

cant desorption from Pt_oxsites occurring only above 200 °C and

incomplete even at 350 °C, Figure S6(B). The stronger adsorption

of CO to Pt_oxclusters, compared to Pt_metalclusters, is consistent with

previous results on Pt single crystals and supported Pt.^45,47,48,66Dur-

ing CO desorption from Pt_oxsites various shifts in the CO band

frequency and width were observed, suggestive of evolving oxida-

tion states and structure of adsorption sites during the temperature

ramp.

With an understanding of the comparative strength of interaction

of CO with the various adsorption sites, we also wanted to analyze

the relative reactivity of CO bound to these sites in transient exper-

iments. We focused on CO bound to Pt_metaland Pt_oxsites, because

CO binds too weakly to Pt_isosites to analyze reactivity in transient

experiments. Figure 5(B) and (C) show IR spectra collected on

respectively reduced and oxidized 1 wt% DI catalysts, during the

following experimental protocol: CO adsorption to saturation at

room temperature, He exposure at room temperature for 10

minutes, a 10% O₂/ He mixture for 10 minutes, and a temperature

programmed oxidation (TPO) ramp (β = 10 °C/min) to 500 °C in

the 10% O₂mixture. It is seen in Figure 5(B) that CO adsorbed to

Pt_metalsites is stable in the presence of He at room temperature,

consistent with Figure S6(A). Upon exposure to the 10% O₂/He

mixture at room temperature, CO reacts to produce CO₂.³³During

the TPO, the remaining CO is converted to CO₂almost completely

by 120 °C, with only a small amount of CO adsorbed to Pt_oxsites

(formed during the oxidation ramp) remaining intact until above

300 °C. This shows the high reactivity of CO on Pt_metalsites for

oxidation by O₂, as is well known.

ACS Paragon Plus Environment

Page 9 of 17

Journal of the American Chemical Society

Conversely, in Figure 5(C) it was observed that CO bound to Pt_ox

110.0E-02

1E-3

100

1

2

3

4

5

6

7

8

0.025%

1%

0.025%

2

O₂

sites on the pre-oxidized 1 wt% DI catalyst is significantly less reac-

tive than CO bound to Pt_metalsites observed in Figure 5(B), with

minimal reactivity until > 160 °C. As in Figure 4(A), initially the

pre-oxidized catalyst has some metallic character due to reduction

of Pt_oxclusters by CO, as evidenced by the CO stretching frequen-

cies characteristic of adsorption on Pt_metalsites (2040-2090 cm^-1).

The difference in reactivity of CO bound to Pt_oxand Pt_metalsites is

highlighted in Figure 5(C) by the almost complete loss in CO

bound to Pt_metalsites following O₂introduction at room tempera-

ture, with the retained existence of CO bound to Pt_oxsites until

much higher temperature.

(B)

(A)

0.025% CO

1%O₂2

Order = 0.51

1% CO

1.E-3

100

y = -68.5x + 9.92

1E1-04

Order = 0.01

10

1.E-4

Order = 0.52

Order = 0.04

y = -52.7x + 4.14

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.E-5

1

1E-5

0.001

0.25

0.26

1/RT (kJ/mol)^-1

0.27

0.28

0.01

P_CO,O2(atm)

0.1

The TPD and TPO experiments on Pt_iso, Pt_metaland Pt_oxsites high-

light several unique characteristics distinguishing site-specific spec-

troscopic and reactive characteristics of adsorbed CO. First, CO

adsorption strength is in the order Pt_iso<<Pt_metal<Pt_ox. Second, CO

adsorbed to the Pt_isospecies exhibit constant FWHM, band posi-

tion, and band symmetry with changing coverage, which is in stark

contrast to observations on Pt_oxand Pt_metalsites, demonstrating the

homogeneity of the produced Pt_isospecies and further substantiat-

ing the IR assignment.⁶⁸Third, CO adsorbed to Pt_oxsites is signifi-

cantly less reactive in O₂than CO adsorbed to Pt_metalsites, suggest-

ing Pt_oxsites play no role in steady state CO oxidation reactivity

measurements on Pt/TiO₂catalysts.

Figure 6. CO oxidation Kinetics. (A). Arrhenius plots showing

temperature dependence of the per-gram Pt rate (mol·s^-1·g_Pt^-1) of CO

oxidation on each catalyst from 160-200 °C in a 200 sccm flow of 1%

-

CO, 1% O₂, balance He. (B) Linearized per-gram Pt rates (mol·s^-1·g_Pt

¹) as a function of CO and O₂pressure. The reported data is an aver-

age of three unique sets of measurements on freshly loaded catalysts

into the reactor, ensuring reproducibility.

The catalysts were reduced for 2 hours in 20 sccm of pure H₂at 300

°C to ensure all clusters were Pt_metaland then were exposed to 1%

CO, 1% O₂, 98% He at 200 °C for 10 hours to reach steady state

reactivity. Figure 6(A) shows the rate of CO oxidation as a function

of temperature over the two catalysts, where it was observed that

the apparent activation energy (E_app) on the Pt_isospecies is higher

than on the 1 nm Pt_metalclusters (69 kJ/mol on Pt_isoversus 53

kJ/mol on Pt_metalsites). The measured E_appfor both catalysts are

consistent with previous measurements for CO oxidation over Pt

3.3 Steady-state CO oxidation reactivity of Pt_iso

and Pt_metal

To directly compare the reactivity of Pt_isoand Pt_metalsites for the

CO oxidation reaction, steady state measurements were executed

under strict kinetic control as a function of temperature and reac-

tant partial pressure. 1 wt% DI and 0.025 wt% SEA catalysts were

used to compare the reactivity of these sites, as the 1 wt% DI mate-

rial contains almost exclusively 1 nm Pt clusters, while the 0.025

wt% SEA material contained almost exclusively Pt_isospecies. To

minimize the partial pressure drop across the catalyst bed, hot spot

formation, and mass transfer limitations, the catalysts were co-

impregnated onto a SiO₂gel support and further diluted in SiO₂, as

described in the methods section. Flow rate dependent kinetic

measurements provide evidence of kinetic control in the rate meas-

urements, see Figure S1.

69

on reducible supports, which range from 40-70 kJ/mol on CeO₂

and 50-60 kJ/mol on TiO₂.⁷⁰Higher barriers around 85 kJ/mol are

typically observed when Pt is supported on non-reducible supports

like Al₂O₃.^52,71Pt_isospecies exhibited enhanced CO oxidation rates

on a per gram Pt basis as compared to 1 nm Pt_metalclusters, ranging

from 4x greater at 140 °C to 6x greater at 200 °C. Following CO

chemisorption measurements for the 1 wt% DI catalyst to deter-

mine the number of available Pt sites, and assuming 100% disper-

sion for the 0.025 wt% SEA catalyst, rates were converted to turn

over frequencies (TOF), see Figure S8. At 140 °C, the TOF on Pt_iso

sites is ~12% higher than on 1 nm Pt_metalclusters, while at 200 °C

the TOF is ~2x greater on Pt_isocompared to 1 nm Pt_metalclusters.

ACS Paragon Plus Environment

Journal of the American Chemical Society

Page 10 of 17

1

2

3

4

5

6

7

8

0.005

0.004

0.003

0.002

0.001

0.000

0.005

0.004

0.003

0.002

0.001

0.000

0.005

CO₂

He 2.5min

He 3min

He 4min

He 5min

He 7min

He 8min

He 10min

He 2.5min

He 3min

He 4min

He 5min

He 7min

He 8min

He 10min

(A)

(B)

(C)

(g)

0.004

0.003

0.002

0.001

0.000

CO-Pt_iso

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

CO_(g)

CO-Pt_metal

2050

CO-Pt_metal

2100 2000

CO-Pt_metal

2100 2000

2500

2350

2200

1900

2200

1900

2200

1900

Wavenumber (cm^-1)

0.004

0.003

0.002

0.001

0.000

0.0015

0.0010

0.0005

0.0000

0.0015

0.0010

0.0005

0.0000

CO-Pt_metal,UC

(D)

(E)

(F)

CO-Pt_metal,UC

CO-Pt_metal,WC

CO-Pt_metal,UC

CO-Pt_metal,WC

CO-Pt_ox

CO₂

(g)

CO_(g)

2200

2100

2000

1900

2500

2350

2200

2050

1900

2200

2100

2000

1900

Wavenumber (cm^-1)

Figure 7. Catalyst stability under reaction conditions and in-situ analysis. IR spectrum of CO adsorbed at room temperature and satu-

ration coverage to the 0.025 wt% Pt/TiO₂SEA catalyst during a He flush prior to reaction, (A), and after 64 hours on stream, (B). (C) In-

situ IR spectrum collected at 200 °C after 2hr of CO Oxidation (1% CO, 1% O₂, balance He) from the 0.025 wt% Pt/TiO₂SEA catalyst. IR

spectrum of CO adsorbed at room temperature and saturation coverage to the 1 wt% DI catalyst during a He flush prior to reaction, (D),

and after 64 hours on stream, (E). The band centered at 2080 cm^-1is assigned to CO on well-coordinated Pt_metalsites, CO-Pt_metal,WC, while

the band centered at 2060 cm^-1is assigned to CO on under-coordinated Pt_metalsites, CO-Pt_metal,UC. (F) In-situ IR spectrum collected at 200

°C after 2hr of CO Oxidation (1% CO, 1% O₂, balance He) from the 1 wt% Pt/TiO₂SEA catalyst.

Partial pressure dependence measurements were made on the two

catalysts at 200 °C, as shown in Figure 6(B). From the slopes of the

linearized plots in Figure 6(B), we found that the reaction rate was

not dependent on CO partial pressure and dependent on O₂partial

pressure with a ½ order dependence, and thus we derived a con-

readily desorbs at room temperature. This is convincing evidence

that the predominant Pt_isostructure is retained even after ~64 hours

on stream under CO oxidation conditions. Furthermore, to sub-

stantiate that kinetic measurements were not skewed by structural

changes to the catalyst as it aged on the shelf between repetitive

measurements, the 0.025 wt% SEA catalyst was characterized by

CO probe molecule IR one month after initial synthesis and

showed no change in the reported signatures CO on Pt_isosites, see

Figure S9, further substantiating the excellent stability of Pt_isospe-

cies in the catalyst architectures used in this study.

!/!

69,72

sistent rate law on both catalysts of, ꢀꢁꢂꢃ ∝ ꢀ

.

The differ-

!

ence in rates between the 1 nm Pt_metalclusters ^!and Pt_isospecies in

the partial pressure dependent measurements was comparable to

the difference measured at 200 °C in the temperature dependent

studies (Figure 6(A)), with Pt_isospecies being 6-8x more reactive

on a per g Pt basis. To ensure that the Pt_isoand 1 nm Pt_metalclusters

retained their structures following the ~64 hours on stream under

CO oxidation conditions that was required to obtain kinetic data,

the two catalysts were characterized by CO probe molecule IR pre-

and post-reaction. In Figure 7(A) and 7(B), the IR spectra associ-

ated with CO adsorbed to the 0.025 wt% SEA catalyst pre- and

post-reaction is shown as a function of time in a He flow at room

temperature. The pre- and post-reaction catalysts show almost

identical spectra, with an 8 cm^-1FWHM for CO adsorbed to Pt_isoat

2112 cm^-1that displays a constant frequency and FWHM as CO

In Figure 7(D) and (F), the IR spectra of CO adsorbed to the re-

duced pre- and post-reaction 1% DI catalyst at saturation coverage

and room temperature are shown. Some small changes in the shape

of the CO IR spectrum in the post-reaction catalyst is evident, sug-

gesting some restructuring of the Pt catalyst due to exposure to CO

oxidation conditions, which has been reported previously.^52,73

However, the reconstruction of Pt particles induced by CO oxida-

tion conditions is minimal and known to occur on very rapid time

scales prior to steady-state reactivity being reached, thus recon-

struction did not influence the kinetic comparison.

ACS Paragon Plus Environment

Page 11 of 17

Journal of the American Chemical Society

To obtain insights into the operating state of Pt_isoand Pt_metalclusters Even in these cases, overlapping TiO₂particles in the images could

1

2

3

4

5

6

7

8

under CO oxidation conditions, in-situ IR spectra were taken at

reaction conditions (1% CO, 1% O₂, 98% He at 200 °C) over a 2

hour period at which point subsequent spectra were identical, indi-

cating steady state CO oxidation had been achieved. In Figure

7(C), a representative in-situ IR spectrum collected from the 0.025

wt% SEA catalyst spanning the CO and CO₂IR stretching frequen-

cies is shown. Bands in the range of 2250-2400 cm^-1shows that

CO₂production is occurring, while the bands between 2150-2200

cm^-1and 2050-2150 cm^-1are associated with gas phase CO vibra-

tions. Interestingly, there is no visible stretch associated with CO

adsorbed to Pt_isosites under steady state reaction conditions. A

small shoulder around 2060 cm^-1is present, which is likely due to

CO adsorbed to the minority of sub-nm Pt clusters present on the

0.025% SEA catalysts. The in-situ spectrum collected from the 1

wt% DI catalyst shows the presence of gas phase CO₂and CO,

consistent with the 0.025 wt% catalyst, but two bands at 2080 and

2060 cm^-1are also present, suggesting CO is adsorbed to Pt_metalsites

at significant coverage under reaction conditions. Overlapping with

the gas phase CO is a broad band associated with CO adsorbed to a

cationic Pt site at 2120 cm^-1. The in-situ IR spectra provide an in-

teresting insight that while the Pt_isoand 1 nm Pt_metalcluster active

sites exhibit very similar rate laws, the coverage of CO on these sites

under reaction conditions is quite different. The mechanistic impli-

cations of this result will be discussed below.

not be ruled out as a source of close Pt_isospecies proximity. Thus,

the STEM imaging supported the hypothesis that primarily 1 Pt

atom was deposited per TiO₂particle.

This conclusion agrees with the observed trends in the influence of

synthetic conditions on the production of Pt_isospecies versus Pt

clusters support this conclusion. For example, for catalysts synthe-

sized at weight loadings consistent with having less than 1 Pt atom

per TiO₂particle (< 0.1 wt%), the IR signature of CO on Pt_isospe-

cies dominated, provided that the synthesis was executed under

extremely dilute and homogeneously mixed conditions where the

interaction between Pt ions and TiO₂particles in solution would be

rare and thus the deposition of a single Pt ion per TiO₂particle

would be favored. Furthermore, weight loading dependent meas-

urements shown in Figure S2 indicate that even when using opti-

mized synthetic conditions, if the Pt weight loading dictated that on

average there should be greater than 1 Pt atom per TiO₂particle

then Pt cluster formation occurred. Together these considerations

provide strong evidence that our synthetic strategy induced the

deposition of ~1 Pt atom per TiO₂particle when operating at suffi-

ciently low weight loadings and that this architecture was the pri-

mary reason the Pt_isospecies exhibited excellent stability. The high

stability was critical to differentiate the spectroscopic signatures

and reactivity of the Pt_isospecies from Pt_metaland Pt_oxclusters.

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

We next focus on the IR spectra and adsorption energy of CO on

the various Pt sites. CO adsorption on Pt surfaces, nanoparticles,

and clusters has been studied in great detail using IR and TPD and

our results on Pt_metaland Pt_oxclusters agree well with previous re-

ports. CO adsorbed to 1 nm Pt_metalclusters on the 1 wt% DI catalyst

exhibited two distinct bands centered at 2077 cm^-1and 2058 cm^-1

respectively assigned to CO adsorbed to WC and UC Pt atoms on

the clusters.^52,58,61,62Greater intensity was observed for the band

associated with CO adsorption on UC sites, which is consistent

with the expected higher concentration of UC sites compared to

WC sites given the distribution of Pt cluster and particle sizes ob-

served by STEM (1 nm diameter clusters with some 4 nm parti-

cles). On the pre-oxidized 1 wt% DI catalyst, the IR spectrum of

adsorbed CO exhibited a broad band at 2118 cm^-1that was as-

signed to CO adsorption on Pt_oxclusters.⁶⁴This assignment has

been made previously and was substantiated here by STEM analy-

sis showing constant Pt particle size regardless of reduction or oxi-

dation pretreatment, Figure S5. The blueshifted stretching fre-

quency of CO adsorbed to Pt_oxclusters, compared to Pt_metalclusters

of the same size, derives from the decreased charge transfer to CO,

due to the cationic oxidation state of Pt in Pt_oxclusters. All of the IR

analysis is consistent with previous results.^40,45,64

4. Discussion

Compared to previous reports where Pt_isospecies were observed to

form Pt clusters under oxidative conditions, reductive conditions,

reaction conditions, or under an electron beam in a TEM, the Pt_iso

species reported here on TiO₂nanoparticles retained their struc-

ture and homogeneous distribution under all considered condi-

tions.^33,34,37The lack of stability for most Pt_isospecies under various

conditions is not surprising, given the known mechanism of Pt

sintering through Pt atom emission and migration on supports.⁷⁴

One approach to overcome the low stability of Pt_isospecies is to

provide anchoring sites, for example coordinatively unsaturated

Al³⁺sites or square planar binding pockets on CeO₂, where Pt_iso

species bind particularly strongly.^8,32,75,76However, even in these

cases there is no evidence that Pt_isospecies remain stable under

reduction conditions, which are required to differentiate the reac-

tivity of Pt_isospecies and Pt_oxclusters.

An alternative approach, which was proposed here, is to isolate a

single Pt atom on a single support particle – effectively marooning

Pt_isospecies on TiO₂islands – such that Pt aggregation would re-

quire an energetically unfavorable hop between support particles.⁷⁷

This idea has been proposed recently, where phosphotungstic or

phosphomolybdic acids have been used to host single Pt or Rh

atoms.^78,79These small acidic clusters are interesting supports for

single metal atoms because they are well defined and small. How-

ever, the stability of these catalysts is limited by the low stability of

the supports, which degrade below 200 °C, thus limiting their ap-

plicability.

The adsorption energy of CO was observed by TPD to be stronger

on Pt_oxclusters than on Pt_metalclusters, indicated by the higher de-

sorption temperature. This is in line with previous results where the

maximum rate of CO desorption from Pt_oxwas observed to occur at

temperatures 100 °C higher than on Pt_metal(280 °C for Pt_oxversus

180 °C for Pt_metal) during TPD experiments from reduced and oxi-

dized Pt single crystals.^47,48The stronger adsorption energy of CO

on Pt_oxclusters explains why these sites exhibit extremely low reac-

tivity for CO oxidation – CO poisons the catalytic surface. This

also highlights the importance of differentiating Pt_isoand Pt_oxspe-

cies in analysis of reactivity, because without differentiation the

large adsorption energy of CO and low CO oxidation reactivity on

Pt_oxcould confuse the inherent reactivity of Pt_isosites.

The evident stability of the Pt_isospecies reported here under a wide

range of conditions, with no specific anchoring sites on the support,

is suggestive that our synthesis strategy was successful in primarily

depositing a single Pt atom per TiO₂particle. In an analysis of the

STEM imaging of the 0.05% SEA catalysts shown in Figure 3D,

only 2 of the 30 images containing Pt_isospecies showed evidence of

multiple Pt_isospecies per TiO₂particle, see Figure S10 for examples.

ACS Paragon Plus Environment

Journal of the American Chemical Society

Page 12 of 17

While IR spectroscopic analysis of CO bound to Pt_metaland Pt_oxWhile we cannot comment directly on the structure of the TiO₂site

1

2

3

4

5

6

7

8

clusters is abundant in literature, there are far fewer rigorous anal-

yses of CO bound to oxide supported Pt_isospecies. Previous reports

of the vibrational frequency for assigned bands associated with CO

bound to Pt_isospecies vary with changes in the support (Pt_iso/H-

Mordenite,⁴⁰2123 cm^-1; Pt_iso/H-ZSM-5,³³2115 cm^-1; Pt_iso/CeO₂,

2095 cm^-1;³²Pt_iso/FeO_x,²⁷2080 cm^-1), but are generally comparable

to 2112 cm^-1reported here for Pt_iso/TiO₂. The variation in band

position of CO bound to Pt_isospecies is likely caused by varying

interactions of the Pt_isospecies with the support, where previous

studies have found that the stretching frequency of CO bound to

supported metals redshifts when going from more acidic to more

basic supports.⁶⁸Consistent in all reports is the ~40 cm^-1blueshift

of the frequency of the CO stretch when bound to Pt_isospecies

compared to Pt_metalclusters. The common blueshifted CO stretch-

ing frequency on Pt_isospecies derives from the cationic oxidation

state (likely 2+), which stems from direct coordination of Pt_isowith

oxygen atoms on the oxide support. Additionally, similar to a few of

the previous reports, the band associated with CO stretching when

adsorbed to Pt_iso/TiO₂was observed to exhibit a constant frequen-

cy as a function of CO coverage, which occurs due to the lack of

dipole-dipole coupling of adjacent CO ligands on dispersed adsorp-

tion sites, and is a critical spectroscopic signature of CO bound to

isolated metal sites.^27,40,44

where Pt_isopredominantly adsorbed, we can use the observed CO

adsorption strength to infer some characteristics of the site. A very

weak adsorption energy of CO on Pt_iso/TiO₂can be inferred from

the observed facile desorption of CO at room temperature in an

inert atmosphere. This is similar to the predicted weak adsorption

on CO on Pt_isospecies adsorbed in square planar configurations to

CeO₂surfaces, where the Pt_isospecies is calculated to be adsorbed

very strongly to CeO₂.^8,29Thus, it is hypothesized that Pt_isospecies

reported here are bound to a location (or a few similar locations)

on the TiO₂support with a very strong adsorption energy, such that

their strong coordination to the support reduces the binding energy

of CO. The weak Pt-CO interaction for Pt_iso/TiO₂reported here

suggests that CO does not render the Pt atom mobile as has previ-

ously been observed,⁸⁰which is in agreement with the tight FHWM

of the CO stretching frequency that indicates Pt is primarily located

at a single location on the support.

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

A previous report examined the adsorption location of Pt_isospecies

on rutile TiO₂, which suggest O-vacancies are the most stable sites,

although this information cannot be directly related to our work

because the TiO₂support nanoparticles were 100% anatase.⁸¹

However, given the small size and high density of defects that are

expected to exist on 5 nm diameter anatase crystals used as sup-

ports in this study, it would not be surprising if the predominant

adsorption site for Pt_isospecies was a defect or step site on TiO₂. It

should be noted that it is possible that impurities in the TiO₂could

provide an anchoring site for Pt_isospecies, although the quantity of

any single impurity in the TiO₂used here is smaller than the

amount of Pt, suggesting impurities did not provide a stable bind-

ing site. Our results are suggestive that the Pt_isospecies analyzed

here are essentially single site materials, and that Pt_isois adsorbed

strongly enough to the TiO₂support to minimize metal atom mo-

bility, creating a weak interaction with CO.

Two emerging and important questions associated with single met-

al atom catalysts on oxide supports are: (1) Are these single site

catalysts, meaning all metal atoms sit on the same site on the sup-

port, or are metal atoms distributed over many types of sites on the

support? (2) Where on the support the precious metal atoms sit?

It has been proposed that the FWHM of the adsorbed CO stretch-

ing band can be used to qualitatively address the question of how

homogeneously the metal atoms are distributed on a support, with

the idea that a broader CO band FWHM is suggestive of isolated

metal atoms sitting in a variety of local environments, which creates

a range of varying interactions between the metal atom and CO.⁶⁸

Previous reports of IR spectra for CO adsorbed to Pt_isospecies have

shown relatively broad FWHM, ranging from 15 to > 30 cm^-1(Pt_i-

_so/H-Mordenite,⁴⁰~18 cm^-1; Pt_iso/H-ZSM-5,³³~25 cm^-1; Pt_i-

_so/CeO₂,³²30 cm^-1; Pt_iso/FeO_x,²⁷>30 cm^-1), signifying that Pt_isospe-

cies are adsorbed to a range of different sites on the support. In

comparison, we report much tighter FWHM between 6-10 cm^-1for

the CO stretching band when adsorbed to Pt_iso/TiO₂, which is

indicative that the Pt_isospecies are homogeneously distributed at a

single, or a few very similar, adsorption site on the TiO₂support. A

recent analysis of Ir single atoms on carefully prepared zeolite sup-

ports reported a ~5 cm^-1FWHM, which compared well to a 4 cm^-1

FWHM for a homogeneous organometallic Ir carbonyl complex in

solution, revealing that the zeolite supported single Ir atoms is like-

ly a single site material. Comparing the Ir case to the 6 cm^-1FWHM

observed for the CO stretch when adsorbed to Pt_iso/TiO₂observed

here, it seems that our synthetic approach is able to localize the Pt_iso

species at a single adsorption site on TiO₂, although we do not have

a homogeneous Pt-CO analogue for direct comparison to prove

this. Interestingly, it can be seen in Figure 2B that when catalysts

were synthesized with lower pH the CO band broadens (for exam-

ple to a FWHM of 22 cm^-1at a synthesis pH of 8) and loses sym-

metry, suggesting the existence of Pt_isospecies at multiple adsorp-

tion sites on the TiO₂support. It seems that careful control of syn-

thetic conditions are required to deposit Pt_isospecies at consistent

sites on the support.

An interesting apparent paradox that emerges from our results is

the significant difference in CO adsorption energy on Pt_isospecies

and Pt_oxclusters. The Pt_oxclusters form from partial CO reduction

of pre-oxidized Pt_metalclusters, and contain CO adsorption sites that

are characteristic of Pt²⁺and Pt⁰, Figure 4 and 5. We propose that

on these partially oxidized Pt clusters, some Pt²⁺species remain in

under-coordinated geometries, perhaps bound locally to only 2

oxygen atoms, enabling strong binding to CO. Conversely, the

weak binding between Pt_isospecies and CO suggests that these Pt²⁺

species are strongly locally coordinated to oxygen atoms in TiO₂.

We propose that the two Pt species with apparently similar oxida-

tion state (likely Pt²⁺) exhibit strong differences in interaction

strength with CO due to differences in their local coordination.⁸²

These details will be explored more in the future.

Next we discuss the steady-state CO oxidation reactivity of Pt_i-

_so/TiO₂, and comparative reactivity of 1 nm Pt_metalclusters on TiO₂,

in the context of previous reports for Pt_isoon other supports. A pre-

vious report of CO oxidation over Pt/FeO_xcatalysts showed a ~2-

fold increase in CO oxidation TOF for Pt_isosites compared to Pt_metal

sites on 1 nm Pt clusters, which is in excellent agreement with our

results for similar Pt species on TiO₂.²⁷However, the Pt_iso/FeO_x

catalyst exhibited significantly higher CO oxidation TOF as com-

pared to our reported Pt_iso/TiO₂catalyst, where the same TOF

(~0.1 s^-1) observed on Pt_iso/TiO₂at 200 °C was observed at room

temperature for Pt_iso/FeO_x. More recently, it was reported that

Fe₂O₃supported Pt_isoexhibited a CO oxidation TOF that is ~4-fold

higher than Pt_iso/TiO₂reported when correcting for differences in

ACS Paragon Plus Environment

Page 13 of 17

Journal of the American Chemical Society

O₂partial pressure.⁸³ZnO supported Pt_isowas also examined in the

havior of late transition metals promoting the reducibility of ox-

ides.⁸²For example, it was inferred from microkinetic modeling that

under water gas shift reaction conditions on Pt/TiO₂catalysts, a

non-negligible concentration of oxygen vacancies at the Pt/TiO₂

interface could be stable. This is in contrast to the low vacancy

concentration on bare TiO₂under the same conditions.^82,86Experi-

mental studies using isotopic labeling of oxygen have also implicat-

ed oxygen atoms in TiO₂as active species in the water gas shift

reaction on Pt/TiO₂catalysts.⁸⁷

1

2

3

4

5

6

7

8

same report showing almost identical TOFs (when corrected for

differences in O₂partial pressure) as observed here for TiO₂. We

attribute reactivity differences for Pt_isoon these reducible supports

to differences in the activation energy for oxygen abstraction from

the supports, as discussed below.⁸⁴

In contrast to the comparison with Pt_isoreducible supports, Pt_i-

_so/TiO₂exhibited significantly higher reactivity for CO oxidation as

34

compared to Pt_iso/Θ-Al₂O₃and Pt_iso/γ-Al₂O₃.⁸³Pt_iso/TiO₂showed

~10x higher TOF compared to Pt_iso/Θ-Al₂O₃at 200°C (0.013 s^-1vs

.11 s^-1) despite the 3.7x higher oxygen partial pressure in the exper-

iments on Pt_iso/Θ-Al₂O₃, which given our deduced rate law suggest

a 20x higher inherent TOF for CO oxidation on Pt_iso/TiO₂. Similar-

ly, ~4x higher TOF was observed compared to Pt_iso/γ-Al₂O₃despite

4x higher oxygen partial pressure, suggesting roughly 8x higher

inherent TOF for CO oxidation on Pt_iso/TiO₂. A comparison of

CO oxidation on Pt_iso/Θ-Al₂O₃and 1 nm Pt_metalclusters on Θ-Al₂O₃

showed a 5x higher reactivity for Pt_metalclusters. This can be under-

stood by the required activation of oxygen from Θ-Al₂O₃in the

catalytic cycle on Pt_iso, which is energetically difficult, compared to

the known CO oxidation mechanism on Al₂O₃supported Pt clus-

ters where the entire reaction cycle occurs on Pt.^52,71

In addition to evidence that Pt can promote the reducibility of

TiO₂, there is strong evidence of relatively facile reverse oxygen

spillover from oxides to small metal clusters and single atoms. For

example, it has been demonstrated via experiments and theoretical

calculations that reverse oxygen spillover from CeO₂to Pt (or Rh)

is an exothermic process and can occur under mild conditions.^88,89

Furthermore, it has been shown via theoretical calculations that for

small Ru and Ni clusters on TiO₂, reverse oxygen spillover is ener-

getically thermoneutral or exothermic and it was further postulated

that this process would be more exothermic on nanoparticles of

TiO₂, as were used in our experiments, compared to extended TiO₂

surfaces used in the models. Thus, we conclude that oxygen migra-

tion from TiO₂to interfacial Pt species could be both energetically

feasible and the rate limiting step under the explored CO oxidation

conditions consistent with our partial pressure dependent meas-

urements and with precedent from previous analyses of reverse

oxygen spillover and the known ability of metals to promote oxide

reducibility.^90,91

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

Pt_isospecies in a KLTL zeolite were reported to achieve TOFs of

0.0038 s^-1prior to oxidation of tetraamine Pt complexes, and 0.012

s^-1following oxidation of the catalysts when operating at 150 °C in

1% CO, 5% O₂, 94% He at atmospheric pressure.³⁵On Pt_iso/TiO₂at

150 °C and 5x lower O₂partial pressure a TOF of 0.015 s^-1was

measured, which suggests that at the same O₂partial pressure the

Pt_iso/TiO₂catalyst would be ~3x more active than the Pt_iso/KLTL

catalyst. In this case no direct comparison to the reactivity of Pt

clusters was made. We find that the reactivity of Pt_isois highly de-

pendent on the support, and that only for reducible oxide support-

ed Pt_isospecies is the CO oxidation reactivity of Pt_isoenhanced

compared to Pt_metalclusters.

In addition to addressing the half order dependence of rate on O₂

partial pressure, it is important to understand mechanistically how

there could be a lack of dependence on CO partial pressure. The

lack of CO partial pressure dependence requires that neither gas

phase CO nor adsorbed CO be involved in the rate-limiting step

and further that the coverage of CO on the active site is essentially

zero. For the case of Pt_isoactive sites, in-situ IR analysis showed no

measurable coverage of CO on Pt_isospecies, which is consistent

with kinetic assessments based on the observed partial pressure

dependences. However, on the 1 nm Pt_metalclusters, significant CO

coverage was observed under reaction conditions. It is proposed

that interfacial Pt atoms, which are cationic due to coordination

with oxygen atoms at the TiO₂surface, exhibit reduced CO binding

energies compared to metallic Pt atoms that exist on the clusters

away from the interface.^{46,83,87,92–94}The reduced CO binding energy

at interfacial Pt atoms, which are coordinated to both the Pt cluster

and the support, minimizes CO coverage under reaction condi-

tions, enabling a mechanism and rate law that is consistent with

observations on Pt_isospecies.

To interpret the support dependent reactivity of Pt_isoapparent from

literature and the site dependent reactivity of Pt_isoversus 1 nm

Pt_metalspecies we observed, we consider the observed CO oxidation

rate law and activation barriers. Both the Pt_isoand the 1 nm Pt_metal

containing catalysts on TiO2₂exhibited identical rate laws with half

order dependence in O₂partial pressure and zero order depend-

ence in CO partial pressure. This is consistent with CO oxidation

measurements made under similar conditions on CeO₂supported

Rh, Pt, Pd and Ni particles.^69,72We observed via in-situ IR that the

Pt_isospecies had low steady state coverage of CO, while the Pt_metal

sites have significant coverage of CO under reaction conditions.

However, the observation of CO coverage on Pt_metalsites under

reaction conditions does not mean that there is a significant cover-

age of CO on sites at the interface between Pt_metalclusters and the

TiO₂surface, where direct bonding between Pt and oxygen on

TiO₂renders Pt atoms cationic.

The involvement of support oxygen in the CO oxidation catalytic

cycles on Pt_isospecies suggests the support dependent TOF meas-

ured for CO oxidation on Pt_iso(inferred from comparison to litera-

ture) likely stems from the differing activation energy associated

with abstraction of atomic oxygen from the oxide support. This can

be further understood by considering the measured CO oxidation

E_appand rate law compared to previous measurements. CO oxida-

tion by Pt_metalclusters supported on Al₂O₃under similar conditions

as reported here shows an E_appof ~85 kJ/mol, a rate law that is first

order in O₂and negative first order in CO, and further shows signif-

icantly lower TOF than reported here for Pt/TiO₂. These differ-

ences provide evidence that the entire reaction cycle occurs on WC

sites on the Pt_metalclusters and that the Pt_metalsurface is saturated by

CO.^52,71Comparison to the Pt/TiO₂results for both Pt_isoand Pt_metal

These results suggest that the reaction proceeds via a Mars van

Krevlen mechanism on both TiO₂supported Pt structures, where

the rate-limiting step involves the migration of atomic oxygen from

TiO₂onto Pt_isoor interfacial Pt atoms on Pt_metalclusters, which is

known as reverse oxygen spillover.^29,69,85The half order reaction

dependence in O₂partial pressure demonstrates that O₂dissocia-

tion is an unlikely rate limiting step, as this would produce a first

order O₂dependence.^69,72The hypothesis of reverse oxygen spillo-

ver from TiO₂to interfacial Pt atoms occurring in the catalytic cycle

and being the rate-limiting step is consistent with well-known be-

ACS Paragon Plus Environment

Journal of the American Chemical Society

Page 14 of 17

sites (where a different rate law and decreased E_appwere observed) catalysts containing exclusively Pt_isospecies are predicted to be 3x

1

2

3

4

5

6

7

8

strongly suggests that the CO oxidation mechanism is different

under identical conditions for reducible and non-reducible oxide

supported Pt structures, and that the difference stems from the

participation of the support in the catalytic cycle on reducible ox-

ides. It is hypothesized that the lower E_appobserved for 1 nm Pt_metal

clusters compared to Pt_isoon TiO₂(53 versus 68 kJ/mol) is due to

the weaker influence of Pt_isospecies on the reducibility of TiO₂and

as a result the increased barrier for oxygen migration onto Pt sites.

Even with the increased barrier on Pt_iso, the TOF for CO oxidation

is greater than on 1 nm Pt_metalclusters, which indicates that more of

the total exposed sites in the Pt_iso/TiO₂catalyst are active sites for

CO oxidation as compared to on 1 nm Pt_metalclusters.

more active on a per-gram Pt basis compared to 1 nm Pt_metalclus-

ters, whereas a 4-6x difference was measured; similarly Pt_isoare

predicted to have a 1.75x greater TOF than the 1 nm Pt_metalclusters,

whereas a ~1.12-2x difference was measured. The agreement be-

tween the relative measured rates and model predicted values is

quite good considering the calculated rates and TOFs assumed a 1

nm Pt cluster structure while STEM analysis confirmed a few larger

Pt particles existed in the 1 wt% DI catalyst and that there was a

small difference in the activation barrier of the two sites (Pt_isovs

Pt_metal) stemming from their differing ability to induce support re-

duction. Therefore, it is concluded that Pt_iso/TiO₂catalysts are the

most efficient utilization of Pt in CO oxidation when supported on

TiO₂, because every atom is exposed for catalysis and interfaces

with support which provides relatively mobile O atoms for the

catalytic cycle via reverse oxygen spillover and weakens the CO

binding energy.

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

Pt_isoatom

1nm Pt_metalCluster

100% Active Sites

33% Active Sites

Conclusions

In conclusion, we demonstrated a synthetic approach and catalyst

architecture where Pt_isospecies were dispersed on 5 nm diameter

TiO₂particles at a ratio of < 1 Pt atom per TiO₂particle, creating

Pt_isospecies that were stable through a wide range of conditions.

Using a correlated STEM imaging and CO probe molecule IR

characterization approach, distinguishing IR signatures of CO

bound to Pt_iso, Pt_oxclusters and Pt_metalclusters on TiO₂were identi-

fied that allow rapid identification and characterization of each type

of site. It was found that the Pt_isospecies produced through the SEA

synthesis approach were essentially single-site species, inferred

from the tight ~6-8 cm^-1FWHM of the adsorbed CO stretching

band. Using the inferred spectroscopic signatures, the order of CO

Non-Active atoms

Active interfacial atoms

Figure 8. Proposed scheme of the active site of Pt_isoand Pt_metal

clusters. The scheme highlights the proposed active sites (green) for

Pt_isoand Pt_metalspecies in CO oxidation. All Pt_isosites are active, while

on 1 nm Pt_metalclusters, only 33% of total atoms (or about 60% of ex-

posed surface atoms) are active interfacial sites. The blue Pt atoms are

non-active sites, which are saturated with CO under reaction condi-

tions but are proposed not to participate in the reaction.

adsorption energy on these sites was identified to be Pt_iso<< Pt_metal

<

Pt_ox, where adsorbed CO on Pt_oxwas essentially inactive for oxida-

tion of CO below ~200 °C. Rigorous kinetic measurements for

steady state CO oxidation on Pt_isoand 1 nm Pt_metalclusters showed

that the Pt_isosites are inherently more reactive because every Pt

atom on the catalyst interfaces with the support where the reaction

takes place and is exposed to reactants. This work demonstrates

that Pt_isoon certain reducible supports provide the most efficient

metal utilization for CO oxidation and further that the architecture

and synthetic approach proposed here could be generally useful for

probing spectroscopic signatures and reactivity of isolated previous

metal atom catalysts.

Based on the above analysis we propose that the catalytic CO oxi-

dation cycle for both 1 nm Pt_metalclusters and Pt_isospecies involves

reverse oxygen spillover from TiO₂to interfacial Pt atoms in the

rate-limiting step. The interfacial Pt atoms in these structures are

suggested to be well-coordinated to their local environment and

cationic, causing the CO adsorption energy to be relatively weak-

ened mitigating CO poisoning that is well known for metallic Pt

atoms under similar conditions. It is important to point out that if

cationic Pt atoms exist in an under-coordinated local environment

and thus bind CO more strongly, as seen for our Pt_oxcase, these

sites will be poisoned by CO until the oxygen is abstracted to form

metallic sites. Thus, a careful balance between charge on the Pt

atom and coordination to local environment, in addition to exist-

ence at an interface with a reducible oxide that enables reverse oxy-

gen spillover, is requisite for forming the most active sites on reduc-

ible oxide supported Pt structures.

ASSOCIATED CONTENT

Supporting Information. Figures S1-10 and associated discussion are

provided in the supporting information. This material is available free

of charge via the Internet at http://pubs.acs.org.

This hypothesis is consistent with recent work analyzing 2 to 20 nm

Pt, Pd and Ni particles on CeO₂where a model was developed that

related the concentration of interfacial metal atoms to the CO oxi-

dation TOF, but extends this trend down to the limit of a single

metal atom active site. As the fraction of total metal atoms existing

at the interface increases 10x when shrinking clusters from 2 nm

down to a single atom, this significantly benefits metal utilization

efficiency. A comparison of the CO oxidation rate per-gram Pt and

TOF on the Pt_isoand 1 nm Pt_metalclusters on TiO2 support this

conclusion, where a proposed 1 nm Pt cluster structure is shown in

Figure 8. Assuming interfacial Pt atoms are the only active site,

AUTHOR INFORMATION

Corresponding Author

* Correspondence to: christopher@engr.ucr.edu

Notes

The authors declare no competing financial interests.

ACKNOWLEDGMENT

P.C. acknowledges funding from University of California, Riverside,

and National Science Foundation (NSF) CAREER grant number

ACS Paragon Plus Environment

Page 15 of 17

Journal of the American Chemical Society

Casty, G. L.; Serna, P.; Corma, A. J. Am. Chem. Soc. 2016, 138 (48),

15743.

CBET-1554112. XP and GWG acknowledge NSF Grant CBET-

1159240 for partial support. STEM Imaging was performed at the

Irvine Materials Research Institute (IMRI).

1

2

3

4

5

6

7

8

(37)

Peterson, E. J.; DeLaRiva, A. T.; Lin, S.; Johnson, R. S.; Guo, H.;

Miller, J. T.; Hun Kwak, J.; Peden, C. H. F.; Kiefer, B.; Allard, L. F.;

Ribeiro, F. H.; Datye, A. K. Nat. Commun. 2014, 5 (4885), 1.

Camacho Rodrigues, A. C.; Fontes Monteiro, J. L. J. Therm. Anal.

Calorim. 2006, 83 (2), 451.

Munoz-Paez, A.; Koningsberger, D. C. J. Phys. Chem. 1995, 99 (12),

4193.

Zholobenko, V. L.; Lei, G.-D.; Carvill, B. T.; Lerner, B. A.; Sachtler,

W. M. H. J. Chem. Soc. Faraday Trans. 1994, 90, 233.

Sun, S.; Zhang, G.; Gauquelin, N.; Chen, N.; Zhou, J.; Yang, S.;

Chen, W.; Meng, X.; Geng, D.; Banis, M. N.; Li, R.; Ye, S.; Knights,

S.; Botton, G. A.; Sham, T.-K.; Sun, X. Sci. Rep. 2013, 3, 422.

Lamberti, C.; Zecchina, A.; Groppo, E.; Bordiga, S. Chem. Soc. Rev.

2010, 39 (12), 4951.

REFERENCES

(38)

(39)

(40)

(41)

(1)

(2)

Faur Ghenciu, A. Curr. Opin. Solid State Mater. Sci. 2002, 6 (5), 389.

Prins, R.; De Beer, V. H. J.; Somorjai, G. A. Catal. Rev. 1989, 31 (1–

2), 1.

(3)

Burch, R.; Breen, J. P.; Meunier, F. C. Appl. Catal. B Environ. 2002,

39 (4), 283.

9

(4)

(5)

(6)

(7)

(8)

Shelef, M.; McCabe, R. . Catal. Today 2000, 62 (1), 35.

Mudd, G. M. Ore Geol. Rev. 2012, 46, 106.

Hunt, L. B.; Lever, F. M. Platin. Met. Rev. 1969, 13 (4), 126.

Tollefson, J. Nature 2007, 450, 334.

Bruix, A.; Lykhach, Y.; Matolínová, I.; Neitzel, A.; Skála, T.; Tsud, N.;

Vorokhta, M.; Stetsovych, V.; Ševčíková, K.; Mysliveček, J.; Fiala, R.;

Václavů, M.; Prince, K. C.; Bruyère, S.; Potin, V.; Illas, F.; Matolín,

V.; Libuda, J.; Neyman, K. M. Angew. Chemie Int. Ed. 2014, 53 (39),

10525.

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

(42)

(43)

(44)

(45)

(46)

Chakarova, K.; Mihaylov, M.; Hadjiivanov, K. Microporous

Mesoporous Mater. 2005, 81 (1–3), 305.

Stakheev, A. Y.; Shpiro, E. S.; Tkachenko, O. P.; Jaeger, N. I.; Schulz-

Ekloff, G. J. Catal. 1997, 169 (1), 382.

(9)

(10)

Liu, J. ACS Catal. 2017, 7 (1), 34.

Yang, X.-F.; Wang, A.; Qiao, B.; Li, J.; Liu, J.; Zhang, T. Acc. Chem.

Res. 2013, 46 (8), 1740.

Ivanova, E.; Mihaylov, M.; Thibault-Starzyk, F.; Daturi, M.;

Hadjiivanov, K. J. Mol. Catal. A Chem. 2007, 274 (1–2), 179.

Ke, J.; Zhu, W.; Jiang, Y.; Si, R.; Wang, Y.-J.; Li, S.-C.; Jin, C.; Liu, H.;

Song, W.-G.; Yan, C.-H.; Zhang, Y.-W. ACS Catal. 2015, 5 (9), 5164.

McCabe, R. .; Schmidt, L. . Surf. Sci. 1977, 65 (1), 189.

McCabe, R. .; Schmidt, L. . Surf. Sci. 1976, 60 (1), 85.

Asokan, C.; DeRita, L.; Christopher, P. Chinese J. Catal. 2017, 38 (1),

1.

Jiao, L.; Regalbuto, J. R. J. Catal. 2008, 260 (2), 329.

Oudenhuijzen, M. K.; Kooyman, P. J.; Tappel, B.; van Bokhoven, J.

A.; Koningsberger, D. C. J. Catal. 2002, 205 (1), 135.

Kale, M. J.; Christopher, P. ACS Catal. 2016, 6 (8), 5599.

Matsubu, J. C.; Zhang, S.; DeRita, L.; Marinkovic, N. S.; Chen, J. G.;

Graham, G. W.; Pan, X.; Christopher, P. Nat. Chem. 2017, 9 (2), 120.

Schreier, M.; Regalbuto, J. R. J. Catal. 2004, 225 (1), 190.

Miller, J. T.; Schreier, M.; Kropf, A. J.; Regalbuto, J. R. J. Catal. 2004,

225 (1), 203.

(11)

(12)

Liang, S.; Hao, C.; Shi, Y. ChemCatChem 2015, 7 (17), 2559.

Zhang, C.; Chen, L.; Cheng, H.; Zhu, X.; Qi, Z. Catal. Today 2016,

276, 55.

(47)

(48)

(49)

(13)

(14)

(15)

Yardimci, D.; Serna, P.; Gates, B. C. ACS Catal. 2012, 2 (10), 2100.

Lu, J.; Serna, P.; Gates, B. C. ACS Catal. 2011, 1 (11), 1549.

Lu, J.; Serna, P.; Aydin, C.; Browning, N. D.; Gates, B. C. J. Am.

Chem. Soc. 2011, 133 (40), 16186.

(50)

(51)

(16)

(17)

Yang, A. C.; Garland, C. W. J. Phys. Chem. 1957, 61 (11), 1504.

Flytzani-Stephanopoulos, M.; Gates, B. C. Annu. Rev. Chem. Biomol.

Eng. 2012, 3, 545.

(52)

(53)

(18)

(19)

(20)

Fu, Q.; Saltsburg, H.; Flytzani-Stephanopoulos, M. Science. 2003, 301

(5635), 935.

Yang, M.; Allard, L. F.; Flytzani-Stephanopoulos, M. J. Am. Chem.

Soc. 2013, 135 (10), 3768.

Zhai, Y.; Pierre, D.; Si, R.; Deng, W.; Ferrin, P.; Nilekar, A. U.; Peng,

G.; Herron, J. A.; Bell, D. C.; Saltsburg, H.; Mavrikakis, M.; Flytzani-

Stephanopoulos, M. Science. 2010, 329 (5999), 1633.

Lin, J.; Wang, A.; Qiao, B.; Liu, X.; Yang, X.; Wang, X.; Liang, J.; Li, J.;

Liu, J.; Zhang, T. J. Am. Chem. Soc. 2013, 135 (41), 15314.

Matsubu, J. C.; Yang, V. N.; Christopher, P. J. Am. Chem. Soc. 2015,

137 (8), 3076.

(54)

(55)

(56)

Lambert, S.; Job, N.; D’Souza, L.; Pereira, M. F. R.; Pirard, R.;

Heinrichs, B.; Figueiredo, J. L.; Pirard, J.-P.; Regalbuto, J. R. J. Catal.

2009, 261 (1), 23.

(21)

(22)

(57)

(58)

(59)

(60)

(61)

Ryczkowski, J. Catal. Today 2001, 68 (4), 263.

Kappers, M. J.; van der Maas, J. H. Catal. Letters 1991, 10 (5), 365.

Primet, M.; Basset, J.; Mathieu, M.; Prettre, M. 1973, 29, 213.

Ewing, G. E. J. Chem. Phys. 1962, 37 (10), 2250.

Hayden, B. E.; Kretzschmar, K.; Bradshaw, A. M.; Greenler, R. G.

Surf. Sci. 1985, 149 (2–3), 394.

Lundwall, M. J.; McClure, S. M.; Goodman, D. W. J. Phys. Chem. C

2010, 114 (17), 7904.

Barshad, Y.; Zhou, X.; Gulari, E. J. Catal. 1985, 94 (1), 128.

Hadjiivanov, K.; Saint-Just, J.; Che, M.; Tatibouët, J.-M.; Lamotte, J.;

Lavalley, J.-C. J. Chem. Soc. Faraday Trans. 1994, 90 (15), 2277.

Bennett, R. A.; Stone, P.; Bowker, M. Catal. Letters 1999, 59 (2/4),

99.

Hadjiivanov, K. I. J. Chem. Soc. Faraday Trans. 1998, 94 (13), 1901.

Hoffmann, F. M. Surf. Sci. Rep. 1983, 3 (2–3), 109.

Hoffman, A. S.; Fang, C.-Y.; Gates, B. C. J. Phys. Chem. Lett. 2016, 7

(19), 3854.

Cargnello, M.; Doan-Nguyen, V. V. T.; Gordon, T. R.; Diaz, R. E.;

Stach, E. A.; Gorte, R. J.; Fornasiero, P.; Murray, C. B. Science. 2013,

341 (6147), 771.

Bamwenda, G. R.; Tsubota, S.; Nakamura, T.; Haruta, M. Catal.

Letters 1997, 44 (1), 83.

Allian, A. D.; Takanabe, K.; Fujdala, K. L.; Hao, X.; Truex, T. J.; Cai,

J.; Buda, C.; Neurock, M.; Iglesia, E. J. Am. Chem. Soc. 2011, 133

(12), 4498.

(23)

(24)

Serna, P.; Gates, B. C. Angew. Chemie Int. Ed. 2011, 50 (24), 5528.

Korhonen, S. T.; Fickel, D. W.; Lobo, R. F.; Weckhuysen, B. M.;

Beale, A. M. Chem. Commun. 2011, 47 (2), 800.

(62)

(25)

(26)

(27)

(28)

Wilburn, David R.; Bleiwas, D. I. U.S. Geol. Surv. Open-File Rep. 2004,

No. 1224, 1.

Alonso, E.; Field, F. R.; Kirchain, R. E. Environ. Sci. Technol. 2012, 46

(23), 12986.

Qiao, B.; Wang, A.; Yang, X.; Allard, L. F.; Jiang, Z.; Cui, Y.; Liu, J.;

Li, J.; Zhang, T. Nat. Chem. 2011, 3 (8), 634.

Zhou, X.; Yang, W.; Chen, Q.; Geng, Z.; Shao, X.; Li, J.; Wang, Y.;

Dai, D.; Chen, W.; Xu, G.; Yang, X.; Wu, K. J. Phys. Chem. C 2016,

120 (3), 1709.

(63)

(64)

(65)

(66)

(67)

(68)

(29)

(30)

Tang, Y.; Wang, Y.-G.; Li, J. J. Phys. Chem. C 2017, 121 (21), 11281.

Aleksandrov, H. A.; Neyman, K. M.; Hadjiivanov, K. I.; Vayssilov, G.

N. Phys. Chem. Chem. Phys. 2016, 18, 22108.

(69)

(31)

(32)

Gao, H. Appl. Surf. Sci. 2016, 379, 347.

Jones, J.; Xiong, H.; DeLaRiva, A. T.; Peterson, E. J.; Pham, H.;

Challa, S. R.; Qi, G.; Oh, S.; Wiebenga, M. H.; Pereira Hernandez, X.

I.; Wang, Y.; Datye, A. K. Science. 2016, 353 (6295), 150.

Ding, K.; Gulec, A.; Johnson, A. M.; Schweitzer, N. M.; Stucky, G. D.;

Marks, L. D.; Stair, P. C. Science. 2015, 350 (6257), 189.

Moses-DeBusk, M.; Yoon, M.; Allard, L. F.; Mullins, D. R.; Wu, Z.;

Yang, X.; Veith, G.; Stocks, G. M.; Narula, C. K. J. Am. Chem. Soc.

2013, 135 (34), 12634.

(70)

(71)

(33)

(34)

(72)

(73)

Zafiris, G. S.; Gorte, R. J. J. Catal. 1993, 143 (1), 86.

Avanesian, T.; Dai, S.; Kale, M. J.; Graham, G. W.; Pan, X.;

Christopher, P. J. Am. Chem. Soc. 2017, 139 (12), 4551.

Carrillo, C.; Johns, T. R.; Xiong, H.; DeLaRiva, A.; Challa, S. R.;

Goeke, R. S.; Artyushkova, K.; Li, W.; Kim, C. H.; Datye, A. K. J.

Phys. Chem. Lett. 2014, 5 (12), 2089.

(35)

(36)

Kistler, J. D.; Chotigkrai, N.; Xu, P.; Enderle, B.; Praserthdam, P.;

Chen, C.-Y.; Browning, N. D.; Gates, B. C. Angew. Chemie Int. Ed.

2014, 53 (34), 8904.

(74)

Moliner, M.; Gabay, J. E.; Kliewer, C. E.; Carr, R. T.; Guzman, J.;

(75)

Kwak, J. H.; Hu, J.; Mei, D.; Yi, C.-W.; Kim, D. H.; Peden, C. H. F.;

ACS Paragon Plus Environment

Journal of the American Chemical Society

Page 16 of 17

Allard, L. F.; Szanyi, J. Science. 2009, 325 (5948), 1670.

16689.

(76)

(77)

Zhang, Z.; Zhu, Y.; Asakura, H.; Zhang, B.; Zhang, J.; Zhou, M.; Han,

Y.; Tanaka, T.; Wang, A.; Zhang, T.; Yan, N. Nat. Commun. 2017, 8,

16100.

Dai, Y.; Lim, B.; Yang, Y.; Cobley, C. M.; Li, W.; Cho, E. C.; Grayson,

B.; Fanson, P. T.; Campbell, C. T.; Sun, Y.; Xia, Y. Angew. Chemie

2010, 122 (44), 8341.

Zhang, B.; Asakura, H.; Zhang, J.; Zhang, J.; De, S.; Yan, N. Angew.

Chemie Int. Ed. 2016, 55 (29), 8319.

Zhang, B.; Asakura, H.; Yan, N. Ind. Eng. Chem. Res. 2017, 56 (13),

3578.

Parkinson, G. S.; Novotny, Z.; Argentero, G.; Schmid, M.; Pavelec, J.;

Kosak, R.; Blaha, P.; Diebold, U. Nat. Mater. 2013, 12 (8), 724.

Chang, T.-Y.; Tanaka, Y.; Ishikawa, R.; Toyoura, K.; Matsunaga, K.;

Ikuhara, Y.; Shibata, N. Nano Lett. 2014, 14 (1), 134.

Ruiz Puigdollers, A.; Schlexer, P.; Tosoni, S.; Pacchioni, G. ACS

Catal. 2017, 7, 6493.

Lou, Y.; Liu, J. Ind. Eng. Chem. Res. 2017, 56 (24), 6916.

An, K.; Alayoglu, S.; Musselwhite, N.; Plamthottam, S.; Melaet, G.;

Lindeman, A. E.; Somorjai, G. A. J. Am. Chem. Soc. 2013, 135 (44),

(85)

Yang, T.; Fukuda, R.; Hosokawa, S.; Tanaka, T.; Sakaki, S.; Ehara, M.

ChemCatChem 2017, 9 (7), 1222.

Ammal, S. C.; Heyden, A. J. Chem. Phys. 2010, 133 (16), 164703.

Kalamaras, C. M.; Panagiotopoulou, P.; Kondarides, D. I.; Efstathiou,

A. M. J. Catal. 2009, 264 (2), 117.

Bruix, A.; Migani, A.; Vayssilov, G. N.; Neyman, K. M.; Libuda, J.;

Illas, F. Phys. Chem. Chem. Phys. 2011, 13 (23), 11384.

Vayssilov, G. N.; Migani, A.; Neyman, K. J. Phys. Chem. C 2011, 115

(32), 16081.

Tosoni, S.; Chen, H.-Y. T.; Pacchioni, G. Surf. Sci. 2016, 646, 230.

Chen, H.-Y. T.; Tosoni, S.; Pacchioni, G. J. Phys. Chem. C 2015, 119

(20), 10856.

Bruix, A.; Rodriguez, J. A.; Ramírez, P. J.; Senanayake, S. D.; Evans, J.;

Park, J. B.; Stacchiola, D.; Liu, P.; Hrbek, J.; Illas, F. J. Am. Chem. Soc.

2012, 134 (21), 8968.

Ammal, S. C.; Heyden, A. J. Phys. Chem. C 2011, 115 (39), 19246.

Hadjiivanov, K. I.; Vayssilov, G. N. Adv. Catal. 2002, 47, 307.

1

2

3

4

5

6

7

8

(86)

(87)

(88)

(89)

(78)

(79)

(80)

(81)

(82)

(90)

(91)

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

(92)

(93)

(94)

(83)

(84)

ACS Paragon Plus Environment

Page 17 of 17

Journal of the American Chemical Society

1

2

stable isolated

Pt atom

3

4

5

6

7

8

2200

2100

2000

Wavenumber (cm^-1)

1 Pt atom / TiO₂particle

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

17

ACS Paragon Plus Environment

Article Doi

Catalyst Architecture for Stable Single Atom Dispersion Enables Site-Specific Spectroscopic and Reactivity Measurements of CO Adsorbed to Pt Atoms, Oxidized Pt Clusters, and Metallic Pt Clusters on TiO2

DOI: 10.1021/jacs.7b07093

Source and publish data:

Authors:

Article abstract of DOI:10.1021/jacs.7b07093

Full text of DOI:10.1021/jacs.7b07093

Products guided by the article

R&D Labs maybe for 7440-06-4

Relevant to this article

Hot Product