Surface Science and Gas Sensing: Promise and Issues

1
Surface Science and Gas SensingPromise and
Issues

• Techniques (STM, Photoemission, etc.) and good
  connection with theory
• Surface (electronic) structure, reactivity,
  influence of dopants, etc.
• Fundamental insights at the atomic scale

? Ultrahigh Vacuum

• ? Single Crystals

2
SnO2
M. Batzill, U. Diebold, The Surface and
Materials Science of Tin Oxide Prog. Surf. Sci.
79 (2005) 47
3
SnO2
M. Batzill, U. Diebold, The Surface and
Materials Science of Tin Oxide Prog. Surf. Sci.
79 (2005) 47
4
Outline

• Surface Structure of Low-index SnO2 Surfaces
• Key insight Dual valency of Sn (Sn2, Sn4)
• SnO2(101) instead of (110)
• Oxidation/reduction
• Theory helpful but limitations
• Surface chemistry of SnO2(101) benzene, water
• fully reduced surface is very unreactive,
  defects on oxidized surface are very reactive
• Growth of Palladium on SnO2
• very unusual 1D growth, governed by kinetics
• Surface Investigations of Nanobelts
• Work in progress, but promising

5
In collaboration with
Bulat Katsiev, Tulane University Matthias
Batzill Now at University of South Florida,
Tampa Alexander Urban and Bernd
Meyer Ruhr-Universität Bochum, Germany Andrei
Kolmakov Southern Illinois University at
Carbondale


NSF-CHE-010908, PRF-40919-AC5
6
SnO2 has the rutile structure
M. Batzill, K. Katsiev, and U. D. , Surf. Sci. ,
529/3 (2003) 295 M. Batzill, K. Katsiev, J.M.
Burst, U. D., A. M. Chaka, and B. Delley, Phys.
Rev. B, 72 (2005) 165414
7
Rutile (110) Lowest Energy Surface, Largest
Facets on Single Crystal

• The SnO2(110) surface forms a series of
  complicated reconstructions in vacuum.
  de Fresart et al. Appl. Surf. Sci. 11/12 (1982)
  637 Cox et al., Surf. Sci. 224 (1989) 121
• The most-stable reconstruction (4x1) forms
  only after sputtering annealing. Pang et al.
  PRB 62 (2000) R7775, Jones et al, Surf Sci. 376
  (1997)367 Sinner-Hettenbach et al, Surf. Sci.
  477 (2001) 50
• Our STM results do not agree with models that
  assume ordered arrays of O vacancies. A. Atrei
  et al., Surf. Sci. 445 (2001)L223 Oviedo et al.,
  Surf. Sci. 463 (2000) 93

Bulk-terminated (1x1) surface
2c-O2
5c-Sn4
M. Batzill, K. Katsiev, and U. D. , Surf. Sci. ,
529/3 (2003) 295
8
Scanning Tunneling Microscopy of SnO2(110)
reconstructions, prepared by sputtering and
annealing in UHV
?
(a-f) Batzill, et al., , Surf. Sci. , 529/3
(2003) 295 (g, h) Pang et al. PRB 62 (2000) R7775
9
STM of SnO2(110)-1x1
Reconstructions not observed after preparation in
high pressure of oxygen 1x1 surface (-gt
1x2).
10
Valence Band Photoemission of SnO2(110)
Sputtered annealed 700C (4x1)
Oxidized 20 mbar O2
Oxidized annealed
Oxidized 100 mbar O2
11
The SnO2(110) surface is quite complex.Some
surface structures can only be achieved in
artifical UHV environment.Beware of surface
models with ordered arrays of vacancies!
12
SnO2(101)
o Much simpler surface o Removal/addition of
surface oxygen does not lead to
reconstructions o Dual valency of Sn 5s2
5p2 Sn2 Sn4 o SnO2(101) films can be
grown epitaxially long sides of nanobelts have
(101) orientation
13
Two surface terminations for SnO2(101)
Sn (5s25p2) has two stable oxidation states
(2,4) SnO2 and SnO are stable
14
Which surface do we get? This depends on the
oxygen chemical potential of the system.
?O
Sn4
?O
Sn2
K. Reuter and M. Scheffler, Phys.Rev.B 65 (2001)
035406
15
Oxygen terminated surface 10 mbar O2 at RT
Tin terminated surface sputtered and annealed in
UHV
M. Batzill, A.M. Chaka, U. Diebold, Europhysics
Lett. 65 (1) (2004) 61
16
Two surface terminations for SnO2(101)STM
Sn-terminated surface

58 x 58 nm2 1.2 V, 1.6 nA
17
Ultraviolet Photoemission Spectroscopy
Sn4
Temperature
Sn2
M. Batzill, et al. Phys. Rev. B 72 (2005) 165414
18
Reduced SnO2 surface

• Surface state deep in the gap. Very little band
  bending. ??? 1eV
• Very inert!
• Benzene interacts very weakly
  
• M. Batzill et al., Applied Physics Letters, 85
  (2004) 5766
• - Water physisorbs (dissociates at stoichiometric
  surface) M. Batzill, et al., Surf.
  Sci. 600 (2006) L29 J. Phys. C., 18 (2006)
  L129-L134

Sn-5s derived surface state
19
Adsorption of Benzene on SnO2(101)Valence Band
Photoemission
Free molecule
Reduced SnO2(101)
Stoichiometric SnO2(101)
Batzill, Katsiev, Diebold, Applied Physics
Letters, 85 (2004) 5766
20
Water adsorption on oxidized and reduced
SnO2(101) surfaces
Undercoordinated Oxygen may act as Brønsted-base
sitesgt dissociation of water
Fully coordinated Oxygen
Sn2 cations with chemically inert Sn-5s lone
pair.
M. Batzill, et al., Surf. Sci. 600 (2006) L29 J.
Phys. C., 18 (2006) L129-L134
21
UPS of water adsorption on SnO2(101)
T 110 K
reduced surface
22
Water adsorption at 170 K
gt No water adsorption at 170 K on reduced
SnO2(101) surface.
Conclusion The chemical functionality of
SnO2(101) surface depends on its surface
composition.
23
Water induced band bending
24
Surfaces with a SnO-like termination are
structurally simple
25
(No Transcript)
26
Growth of Palladium on SnO2(101)

• 3 dimensional clusters on oxidized surface,
  nucleation at defects
• reduced surface kinetically controlled growth
  of 1-dim clusters

27
Pd on oxidized SnO2(101) STM(all images are 70
nm x 70 nm)
0.06 ML
Clean, oxidized
At oxidized SnO2(101) surface 3D growth
Nucleation at defects
0.12 ML
28
Pd/SnO2(101) STM(all images are 70 nm x 70 nm)
Pd/SnO2
29
Pd growth on reduced SnO2(101)

• most islands are 1 unit cell (5Å) wide and 1 ML
  high
• nucleation at terraces
• length of the islands is limited by terrace size
• neighboring islands rarely merge

Q1 Why do the Pd islands grow longer and not
higher and wider? Q2 Why do they have a specific
width? -gt DFT calculations
60 nm x 60 nm, Vsample 1V, Itunnel 2.2nA
K. Katsiev, et al., Phys. Rev. Lett. 98 (2007)
186102
30
DFT Calculations, Energetics (B. Meyer, A.Urban)
5Å!
Monotonic increase in energy No thermodynamic
driving force to form 1D islands with
a width of 5Å
PBE functional, norm-conserving
pseudopotentials, mixed basis set (PWlocalized
atomic orbitals), 3 layer slabs
31
Annealing of 0.75 ML Pd/SnO2(101) STM (all
images are 70 nm x 70 nm)
Experiment 1D islands are thermally unstable
32
Kinetics!
33
DFT Calculations, Potential Energy Surfaces (B.
Meyer, A.Urban)
Single Pd adatom
-gt Strictly 1D diffusion
34
DFT Calculations, Potential Energy Landscapes (B.
Meyer, A.Urban)
1-Pd stripe
Pd adatom
0.11 eV
- Pd atoms will not stick to the side of the
stripe - stripe will not grow wider
35
DFT Calculations, Potential Energy Landscapes (B.
Meyer, A.Urban)
Pd adatom
1-Pd stripe
0.11 eV
Two 1-Pd stripes
36
DFT Calculations, Potential Energy Landscapes (B.
Meyer, A.Urban)
Pd adatom
1-Pd stripe
0.11 eV
3-Pd stripe
Two 1-Pd stripes
37
A1 1D growth because of strong Pd-Sn
interaction , 1D diffusion, and the absence of
adsorption sites at the sides of stripes.
Q1 Why do the Pd islands grow longer and not
higher and wider? Q2 Why do they have a specific
width?

• Q Why 3-Pd atom stripes?

38
Hypothesis Width of stripe depends on initial
nucleation site!
Importance of Nucleation! -seeding with
nucleation sites?
K. Katsiev, et al., Phys. Rev. Lett. 98 (2007)
186102
39
Width of 1D clusters depends on
initial nucleation site - engineering of wires
with pre-determined width?
Need high-quality SnO2(101) samples!
40
(No Transcript)
41
SEM
TEM
Semiconducting oxide nanobelts
Pan, Dai, Wang Science 291, 1947 (2001).
(100)
(101)
42
SEM of Nanobelts after annealing to 700C in UHV
43
SnO2 - Surface Science

• Surface Structure of Low-index SnO2 Surfaces
• Key insight Dual valency of Sn (Sn2, Sn4)
• SnO2(101) instead of (110)
• Oxidation/reduction
• Theory helpful but limitations
• Surface chemistry of SnO2 Benzene, water
• reduced surface is very unreactive, defects on
  oxidized surface are very reactive
• Growth of Palladium on SnO2
• very unusual 1D growth, governed by kinetics
• Surface Investigations of Nanobelts
• Promising, work in progress