Size-selected Metal Cluster Deposition on Oxide Surfaces: Impact Dynamics and Supported Cluster Chemistry

Sungsik Lee, Masato Aizawa, Chaoyang Fan, Tianpin Wu, and Scott L. AndersonSupport: AFOSR, DOE

CO oxidation on Aun on TiO2

70 atoms

Cluster Size Effects in Catalysis CO oxidation on bulk catalyst

Problems:

TEM characterization

Broad size distribution

Other changes together with size

Bamwenda, ..., Haruta. Catal. Lett. 44, 83, (1997)(as plotted by Goodman et al.

Cluster Size Effects in Catalysis CO oxidation on planar model catalyst

Planar model support / Cluster grown by annealing Au/TiO2

Sized by STM - Considerable size distribution, especially in the small size range.

Valden, Lai, Goodman, Science 281 (1998) 1647

Our size range

Size proportional to metal dose

Metal dose effect on TPD

150 250 350 450 550 650 750 850Temp.(K)

0.05ML0.1ML0.2ML

Ir10 deposited at 1eV/atom

Dose shift is in opposite direction to cluster size shift

Supported catalysts Embedded catalysts

Complete picture of electronic, geometry, and support effects in supported catalysts .

Size selectedMetal clusters

Metal oxide support

What we want to do: Independently vary cluster size, concentration, support structure and defects, morphology of deposit

Characterize all of the above.

Utah Cluster Deposition Instrument

Sputter/Anneal

TPD/TPR

EELSChamber

Load Lock/Anneal

Side View - Main Chamber

STM

XPS

DepositionStage

Cro

ss

R

R

R

Nd YAG Laser532nm/100Hz

Laser Ablation Cluster Source

Quadrupole Ion Guide 2

QMS with Entrance and Exit Lens

Quadrupole Ion Guide 3Focusing Lens

Deposition Stage

Electron Multiplier

Isolation Valve

Quadrupole Ion Guide 1Lens Deflectors

Pressure in main chamber

during deposition2 X 10-9 Torr

9 stagesdifferentialpumping

Mass range~4000 d

0.1ML of Mn(n=1-30)deposited within 30min

Similar source to Heiz

Instrument Design

XPS/ISSAuger

XPS : Ti

Clean TiO2

With Au

Ti3+

Ti4+

O vacancy

O2 interaction with TiO2 vacanciesannealed

sputteredO2

annealed

O2

O2 at vacancies (Gyeong Hwang et al.)

O2 at vacancy

O2 at Ti centers

neighbor 2nd neighbor

Dissociation

XPS : Au 5f~0.3eV shift to higher B.E.

Au4

Au3

Au2

Au

Constant signal = constant sticking coefficient ~1.0

No shifts with cluster size – not seen in any system in this size range

Final StateShift

Initial statechemicalshift

Morphology?

Important: Besenbacher and co. see sintering at room temperature – we should get ~20 atom clusters?

Wahlstrom, E.; Lopez, N.; Schaub, R.; Thostrup, P.; Ronnau, A.; Africh, C.; Laegsgaard, E.; Norskov, J. K.; Besenbacher, F., Phys. Rev. Lett. 2003, 90, 026101/1-01/4.

Morphology: ISS

Ion Beam

Ion Beam

SMSI or EmbeddedClusters

Detection Direction

Metal/Substrate Ratio Near Zero

Substrate signalBlocked/Shadowedin area aroundad-atoms

Ion Beam

SmallClusters

Metal/Substrate Ratio Large

Metal/Substrate Ratio Small

LargeClusters

Embedded SMSI

4He+

1keV 45º incident angle

Only detectatoms intopmostlayer

ISS : cluster size

Au4

Au3

Au2

Au

clean TiO2

Constant intensity: All Au in topmost layer for small clusters?

Are clusters sintering?

scatteringfrom O scattering

from Tiscatteringfrom Au

0 .2 5

0 .2 0

0 .1 5

0 .1 0

0 .0 5

0 .0 0

Met

al /

sub

stra

te ra

tio

1 41 21 08642C lu s te r s ize ( a to m s)

1 .0 0

0 .9 5

0 .9 0

0 .8 5

0 .8 0

(O /

Ti) as

dep

osite

d /

(O /

Ti) cl

eant

io2

I r / su b s tra te O / T i fo r I r n

N i / s u b s t ra te

ISS ratios: Cluster Size Effects

Decrease in O/Ti ratio- Irn preferentially attenuates scattering from O

Anomalous behavior for atoms

ISS probing of CO binding sitesIr

ISS

Inte

nsity

(arb

.)

1.000.960.920.88E/E0

as depositedsputtering 5 minsputtering 10 minsputtering 15 min

Ir10 at 1eV/atom

~70% of COremoved

He+

Aizawa, M.; Lee, S.; Anderson, S. L., Surf. Sci. 2003, 542, 253-75.

Ir IS

S In

tens

ity (a

rb.)

1.000.960.920.88E/E0

Ir10 at 1eV/atomHe+

How to study CO oxidation in UHV?CO sticks to small Au

O2 adsorbs on catalyst

CO2 desorbs at lower temperature than CO – doesn’t poison catalyst

X

All previous UHV studies have used atomic O

ISS : 16O, 18O exchange1. Au3 on Ti16O2

2. 600 L 18O2

3. CO

~20%

18O bound at O vacancies ~9% of surface O is 18O

Some O bound at Au cluster sites – attenuation too low to be atop

Au1/TiO2CO

CO dose effect on ISS: Au1/TiO2

CO on Au

~5 L

CO binds at vacancies – why nothing here?

Au2/TiO2CO

CO dose effect on ISS: Au2/TiO2

CO on Au

Au3/TiO2CO

CO dose effect on ISS: Au3/TiO2

Why No CO on Au?

CO at vacancies

First sign of chemical differences

300 350 400 450 500 550 600 650 700Temp.(K)

5L(28+305L(30)5L(28)0.5L(28+0.5L(30)0.5L(28)0.1L 28

heating rate : 3K/secdose temp. : 300K

~0.1 L C16O exposure

~0.1 L C16Oplus

0.5 L C18O

Red = C16O Blue = C18O Black = all CO

Chemical Probing - TPDTPD: Only probes what desorbs

(CO on Ir2/TiO2)

Inte

nsity

(arb

.)

1.00.80.60.4E/E0

Ir 4f XPS 1keV He+ ISSO

Ti

Ir

Ir2 deposited at 1eV/atom after CO TPD scans

Inte

nsity

(arb

.)

72 70 68 66 64 62 60 58Binding energy (eV)

Bad News

Chemical Probing II

Only probe what desorbs

Only probe what desorbs

Single Crystal Small Clusters

Destructive – 1 shot expt

Less destructive if low T?

TPD:

Pulsed dosing

CO oxidation reaction I : size 600L pre-dose of 18O20.2 L CO pulsesMonitor C16O18O

Au4

Au3

Au2

Au

TiO2

CO oxidation – size effect

Au4

Au3

Au2Au

TiO2

CO oxidation – size effect

Au7 dose ~10 times lower than others

Substantial activity for much smaller clusters than in bulk or model catalyst studies

CO oxidation reaction : Au3decay half width ~ 8 L CO

Decay: O depletion?Clusters changing?

CO pulse reaction II : atom annealing

Growing larger clusters? Unblocking defects?

Au on TiO2

400K

500K

600K

700K

Still much smaller than Au3

CO pulse reaction III : pulse shape

0

50

100

150

200

250

300

350

0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000

time(msec.)

mas

s 46

TiO2 Au1 Au2 Au3 Au4 CO pulse

~300msec

Au4

Au3

Au2

Au

clean TiO2

CO pulsePossible origins of slow CO2 evolution

• slow CO2 thermal desorption rate

• diffusion of active O to reactive site

• lifetime of CO on small Aun

• actual CO oxidation kinetics

What happens with Aun+ are deposited

Expts of Wahlström et al.Should have Aun, n ~ 20 for all samples

CO oxidation & binding strongly dependent on deposited size.

Catalysts are stable on an few hour timescale for many cycles of O2 / CO

Samples do not reach an equilibrium state.

WHY?

Needed• 13CO + 18O2 to allow measurement of both C16O2 and

C16O18O– Only seeing the tip of the iceberg?

• Low temperature deposition and analysis to reduce/change sintering

• Probe steady state species on surface – not just species desorbing

New Endstation

Continuous T control

IRAS capability

Retain load/lock

Greatly simplified manipulations

tip ‘o the hat - Wayne Goodman

Utah Cluster Deposition Instrument

Ti XPSIn

tens

ity (a

rb.)

4 7 0 4 6 5 4 6 0 4 5 5B in d in g e n e rg y (e V )

Ti0Ti3+

Ti4+

clean TiO2

with 0.1ML Ir2, 1eV/atom

Freshly Annealed

1 minute 1 keV He+

~5% missing Odefects

~18% missing O defects

Mg Kα

1253.6 eV

Inte

nsity

(arb

.)

1.00.90.80.70.60.50.40.3 E/E0

Ir

Ir10

Ir2

Ir5

TiO

IrIr15

Irn deposited at 1eV/atom

fig. 4

ISS cluster size effect

Note much lower intensity compared to gold

Ir ISS v.s. Sputter Time

Slow sputtering for large n--Multilayer clusters--Stronger Ir-Ir binding

CO

18+

Inte

nsity

(arb

.)

700600500400Temp. (K)

Ir2

Ir5

Ir10

Ir15

0.1ML Irn at 1eV/atom

clean TiO2

Ir2

clean TiO2

518K

5 L dose of C18O at room temperature

Solid = 1st TPD run

Open = 2nd TPD run

Effects of TPD on the sample

Cluster Size Effects in Catalysis (background)

Example:Gold cluster catalyst, from Haruta, 1997. 2 - 3 nm gold clusters on TiO2

Variation in product yield with cluster size.

Propene/O2/H2 mixture over Au/TiO2 catalyst(from Haruta 1997)