John T. Yates, Jr.John T. Yates, Jr.

Surface Science CenterSurface Science Center

Department of ChemistryDepartment of Chemistry

University of PittsburghUniversity of Pittsburgh

Photochemistry on TiOPhotochemistry on TiO22 Semiconductor Surfaces Semiconductor Surfaces –– How How Kinetic Measurements Can Provide Insight Into The Kinetic Measurements Can Provide Insight Into The

Behavior of Charge CarriersBehavior of Charge Carriers

PIREPIRE--ECCI/ICMR Summer Program onECCI/ICMR Summer Program on

Techniques of Surface Science and CatalysisTechniques of Surface Science and CatalysisUniversity of California University of California –– Santa BarbaraSanta Barbara

August 2006August 2006

1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 20050

200

400

600

800

1000

1200

1400

1600

1800

Num

ber o

f Pub

licat

ions

Year

A. A. FujishimaFujishima and K. Honda, Nature 238 (1972) 37.and K. Honda, Nature 238 (1972) 37.

1972: Water Splitting at n1972: Water Splitting at n--TiOTiO22 ElectrodesElectrodes

www.isiknowledge.com. Search terms = (TiO2OR titanium dioxide AND photo*)

CB

VB+++

hν

3.0 eV

Upon UV excitation, both electrons and holes are photochemically active

towards adsorbates on the TiO2 surface.

DonorMolecule

Product

Product

AcceptorMolecule

Semiconductor Semiconductor PhotoexcitationPhotoexcitation : TiO: TiO22

Sacrificial Hole Scavenger ActionSacrificial Hole Scavenger Action

Solar Spectrum on EarthSolar Spectrum on Earth

Taken from: Fujishima, Hashimoto, Watanabe, “TiO2-Photocatalysis, Fundamentals and Applications”, BKC Inc. Tokyo, 1999

Motivation? TiO2 Based PhotocatalyticTechnology Works!

Motivation? TiO2 Based PhotocatalyticTechnology Works!

TiO2 Based Systems are Efficient Photocatalysts

TiOTiO22 as a Photochemical Substrateas a Photochemical Substrate

2.0nm

5-fold Ti

bridging O vacancy

in-plane O 6-fold Tibridging O

5-fold Ti

[001

]

MezhenneyMezhenney et al. Chemical Physics et al. Chemical Physics Letters, 369 (2003) 152.Letters, 369 (2003) 152.

Atomic Structure of TiOAtomic Structure of TiO22(110) Surface(110) Surface

TiOTiO22(110)(110)--(1x1) surface without oxygen(1x1) surface without oxygenvacancies is vacancies is stochiometricstochiometric (TiO(TiO22))

The surface density of oxygen The surface density of oxygen vacancy sites is regulated by vacancy sites is regulated by annealing conditions (time, annealing conditions (time,

presence of Opresence of O22))

Heating TiOHeating TiO22 above above ~ 700 ~ 700 ––

800 K 800 K →→ OO--vacancy defectsvacancy defects

(reduced TiO(reduced TiO22))

Pan, J. M.; Maschhoff, B. L.; Diebold, U.; Madey, T. E. J. Vac. Sci. Technol. A 1992, 10, 2470.

He Ion Scattering Spectroscopy He Ion Scattering Spectroscopy –– Detects Detects 1818OO22 on Vacancy Defect Siteson Vacancy Defect Sites

∼8% defects in surface adsorb 18O2

Ti: 1s22s22p63s23p63d24s24p0

Ti+4: 1s22s22p63s23p63d04s04p0

R.A. Bartynski, et al.

JVST A10, 2591 (1992)

XPS Evidence for TiOXPS Evidence for TiO22 Reduction by HeatingReduction by Heating

A More Recent Look at A More Recent Look at Electronic Structure of Electronic Structure of Oxygen Defects/TiOOxygen Defects/TiO22(110)(110)

STM STM -- STSSTS

Oxygen Defect StateOxygen Defect State

Spatial distribution of Ti 3d stateSpatial distribution of Ti 3d state

Unoccupied Occupied

Ti5c

T. Minami and M. Kawai T. Minami and M. Kawai (unpublished results)(unpublished results)

Vijay, A.; Mills, G.; Metiu, H. J. Chem. Phys. 2003, 118, 6536.

Use of CO2 Adsorption to Detect Surface Defect Sites Produced

by Heating

Use of COUse of CO2 2 Adsorption to Detect Adsorption to Detect Surface Defect Sites Produced Surface Defect Sites Produced

by Heatingby Heating

Apparatus for Photodesorption Studies from TiO2(110)

°

Measurement Capabilities:Measurement Capabilities:

••Auger Electron SpectroscopyAuger Electron Spectroscopy

••Temperature Programmed Temperature Programmed Desorption (TPD)Desorption (TPD)

••PhotoPhoto--induced Desorption induced Desorption (PD)(PD)

UHV Studies of Photochemical Processes on TiOUHV Studies of Photochemical Processes on TiO22(110)(110)

Comparison of Experimental and Simulated 13CO2 Desorption Spectra on TiO2(110) –

Fully Oxidized Surface

Comparison of Experimental and Simulated Comparison of Experimental and Simulated 1313COCO22 Desorption Spectra on TiODesorption Spectra on TiO22(110) (110) ––

Fully Oxidized SurfaceFully Oxidized Surface

100 150 200 250 300

Experimental Data Simulated Data

Edes = 48.5 kJ/molEinteraction = -10.0kJ/mol

vο = 4.0 x1013 s-1

β =dT/dt =2.0K/sec

Temperature (K)

dθ/dt = νο (1/β) θ exp (-(Ed+Eintθ)/kT)

Δ PCO2

100 150 200 250 3000.0

1.0x10-7

2.0x10-7

3.0x10-7

4.0x10-7

5.0x10-7

6.0x10-7

7.0x10-7

8.0x10-7

QM

S S

igna

l (M

AS

S 4

5)

Temperature (K)

100 150 200 250 3000.0

1.0x10-7

2.0x10-7

3.0x10-7

4.0x10-7

5.0x10-7

6.0x10-7

7.0x10-7

8.0x10-7

QM

S S

igna

l (M

AS

S 4

5)

Temperature (K)

100 150 200 250 3000.0

1.0x10-7

2.0x10-7

3.0x10-7

4.0x10-7

5.0x10-7

6.0x10-7

7.0x10-7

8.0x10-7

QM

S S

igna

l (M

AS

S 4

5)

Temperature (K)

100 150 200 250 300

0.0

1.0x10-7

2.0x10-7

3.0x10-7

4.0x10-7

5.0x10-7

6.0x10-7

7.0x10-7

8.0x10-7

QM

S S

igna

l (M

AS

S 4

5)

Temperature (K)

100 150 200 250 3000.0

1.0x10-7

2.0x10-7

3.0x10-7

4.0x10-7

5.0x10-7

6.0x10-7

7.0x10-7

8.0x10-7

QM

S S

igna

l (M

ASS

45)

Temperature (K)

100 150 200 250 3000.0

1.0x10-7

2.0x10-7

3.0x10-7

4.0x10-7

5.0x10-7

6.0x10-7

7.0x10-7

8.0x10-7

QM

S S

igna

l (M

ASS

45)

Temperature (K)

Successive Reduction of TiO2(110) Using CO2 to Detect Vacancy Defect SitesSuccessive Reduction of TiOSuccessive Reduction of TiO22(110) Using CO(110) Using CO22 to Detect Vacancy Defect Sitesto Detect Vacancy Defect Sites

600K anneal

700K anneal

800K anneal

900K anneal

1000K anneal

1100K anneal

defects

defects

defects

defects

T. L. Thompson, O. T. L. Thompson, O. DiwaldDiwald and J. T. Yates, Jr.and J. T. Yates, Jr.J. Phys. Chem. BJ. Phys. Chem. B 107 (2003) 11700.107 (2003) 11700.

A Recent Controversy About A Recent Controversy About Vacancy Defect Sites on TiOVacancy Defect Sites on TiO22(110)(110)

OxygenOxygen--Mediated Vacancy Mediated Vacancy Diffusion on TiODiffusion on TiO22

2.0nm

STM Image of TiOSTM Image of TiO22(110) Showing Oxygen (110) Showing Oxygen Vacancy Defect Sites Vacancy Defect Sites

Mezhenny, Maksymovych, Thompson, Diwald, Stahl, Walck and Yates. Chem. Phys. Lett. 369 (2003) 152.

Schematic of O2 - Mediated Vacancy Diffusion - TiO2(110)

O

Ti

O

1.

2.

3.

OOO

O

O

Ti

Ti

Ti

Ti

Ti

Ti

O

OTi

OOOO

O

O

Ti

Ti

Ti

Ti

Ti

Ti

O

O

Ti

OOOO

O

O

Ti

Ti

Ti

Ti

Ti

Ti

O

O

O

0 25 50

hν on

ΔIO2

16O18O (34 amu)

18O2 (36 amu)

T = 100 K

2 x 10-8 Amps

t (sec)0 25 50

hν on16O18O (34 amu)

18O2 (36 amu)

T = 180 K

2 x 10-8 Amps

ΔIO2

t (sec)

Isotopomer O2 Photodesorption from TiO2(110) Containing Vacancy Sites

T. Thompson, O. Diwald, J. T. Yates, Jr. Chem. Phys. Lett. 393, 28, 2004.

100 200 300 400 500

18O2 (36 amu)

16O18O (34 amu)

2 x 10-9 AdT/dt = 4 K/sec

ΔIO2

Tads = 105K

T (K)

Isotopomer O2 Thermal Desorptionfrom TiO2(110) Containing Vacancy Sites

T. Thompson, O. Diwald, J. T. Yates, Jr. Chem. Phys. Lett. 393, 28 (2004).

SummarySummary

1. O2-Mediated vacancy diffusion on TiO2(110) is observed by STM.

2. The mechanism proposed is incorrect since isotopic exchange does not occur from lattice oxygen to molecular oxygen.

3. Single Question: What other O2-mediated models can be imagined?

The Controversy Has Been SettledThe Controversy Has Been Settled

•• Bright images were not OBright images were not O22/defect /defect site.site.

•• They were TiThey were Ti--OH groups due to OH groups due to HH22O impurity.O impurity.

Bikondoa, O.; Pang, C. L.; Ithnin, R.; Muryn, C. A.; Onishi, H.; Thornton, G. Nature Materials 2006, 5, 189.

vac – OH + e TiO + H

Detailed Studies of ODetailed Studies of O22PhotodesorptionPhotodesorption from TiOfrom TiO22(110)(110)

•• Providing insight into the mechanisms Providing insight into the mechanisms of photonof photon--induced electroninduced electron--hole pair hole pair production and the activation of production and the activation of adsorbed molecules by these charge adsorbed molecules by these charge carriers.carriers.

a)

b)

Oxygen Vacancies

h+ e-c)

hν Oxygen Photodesorption

--OO

OO

OO

OO

a)

b)

Oxygen Vacancies

h+ e-h+ e-c)

hν Oxygen Photodesorption

--OOOO

OOOO

OOOO

OOOO

Mechanism for Oxygen Mechanism for Oxygen PhotodesorptionPhotodesorption from TiOfrom TiO22(110)(110)

h+ + O2-→ O2↑

•de Lara-Castells, M. P.; Krause, J. L. J. Chem. Phys. 2003, 118, 5098-5105.

OOVACVAC + O+ O22 →→ OO22--

••de Larade Lara--CastellsCastells, M. P.; Krause, J. L. , M. P.; Krause, J. L. J. J. Chem. Phys.Chem. Phys. 2001, 2001, 115115, 4798, 4798--48104810

••de Larade Lara--CastellsCastells, M. P.; Krause, J. L. , M. P.; Krause, J. L. Chem. Phys. Chem. Phys. LettLett.. 2002, 2002, 354354, 483, 483--490.490.

••AnpoAnpo, M.; , M.; CheChe, M.; , M.; FubiniFubini, B.; , B.; GarroneGarrone, E.; , E.; GiamelloGiamello, E.; Paganini, M. , E.; Paganini, M.

C. C. TopTop. Catal. 1999, 8, 189-198.

G. Lu, A. Linsebigler and J. T. Yates, Jr., "The Adsorption and Photodesorption of Oxygen on the TiO2(110) Surface," J. Chem. Phys. 102, 4657 (1995).

QMSHg Lamp

andFilters

Photodiode Detector

QuartzReflector

Fhν0.096 Fhν

Ta

TiO2

OO22 PhotodesorptionPhotodesorption from TiOfrom TiO22(110)(110)

0 10 20 30 40-2

0

2

4

6

8

10

12

14

Y

(A x

10-9

)

Time (s)

Fhν = 4.08 x 1014 photons cm-2s-1

Ehν = 3.4 +/- 0.05 eV

T = 110 K

O2

hυ on

Tracy L. Thompson and John T. Yates, Jr. J. Phys. Chem. B, 109 (2005) 18230.

OO22 PhotodesorptionPhotodesorption from TiOfrom TiO22(110):(110):kk11 FFhhνν

hhνν + TiO+ TiO2 2 →→ e + h e + h kk22

h + T h + T →→ T+ T+ (hole capture by a hole trap)(hole capture by a hole trap)kk33

e + h e + h →→ heat heat (on recombination sites) (on recombination sites) kk44

h + Oh + O22--(a) (a) →→ OO22(g) (g) ↑↑

Rate of Rate of photodesorbingphotodesorbing oxygen scales oxygen scales proportionally with the square root of the incident proportionally with the square root of the incident

light intensity:light intensity:

][][

2

2 2/12/1

32

14 Oh

O Fkk

kkdt

dθ

θυ⎟⎟

⎠

⎞⎜⎜⎝

⎛+

−=

0.0 5.0x106 1.0x107 1.5x107 2.0x107

0

2

4

6

8

10

12

14

hν = 3.4 +/- 0.05 eVT = 110 K

Flux 1/2hν

(photons1/2cm-1sec-1/2)

Y0 (O )

(A

mps

x 1

0-9)

A

B

F1/2hν

(crit.)2

18O2 Initial Photodesorption Yield – TiO2(110)

-10 0 10 20 30 40 50-2

0

2

4

6

8

10

12

14

Ion

Cur

rent

(36

amu)

Time (s)

Fhν = 4.08 x 1014 cm-2s-1

Y (O2)

dhν

Fhν

O2- O2

-O2- O2

-O2-O2

-O2-

TiO2

O2-

Hole trap centers (T)and average range of photogenerated holes which will be trapped.

* * * * * * **

Bulk Hole Trapping Sites Within TiOBulk Hole Trapping Sites Within TiO22

After hole trap filling, k2 ceases to contribute

to the charge exchange process.

Use of Adsorbed Hole Trap Use of Adsorbed Hole Trap Molecules to Increase Trap Molecules to Increase Trap Density Above Bulk DensityDensity Above Bulk Density

Role of CHRole of CH33OH Surface Hole Traps OH Surface Hole Traps on Oon O22 PhotodesorptionPhotodesorption YieldYield

200 300 400 500 600-0.2

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

QM

S Si

gnal

Temperature (K)

TPD of Methanol from TiO2(110)

ab

a

bTiO2

CH 3OH

O 2

0.0 5.0x106 1.0x107 1.5x107 2.0x107

0

2

4

6

8

10

12

14

θ (CH3OH) = 0.6 ML

θ (CH3OH) = 0.2 ML

Flux1/2hυ

(photons1/2cm-1sec-1/2)

hν = 3.4 +/- 0.05 eVT = 110 K

Y

0 (O )

(A

mps

x 1

0-9)

2

θ (CH3OH) = 0 ML

0 1x106 2x106 3x106 4x106 5x106

0.00

0.25

0.50

0.75

1.00

1.25

θ (CH3OH) = .6 ML

θ (CH3OH) = .2 ML

Flux1/2hυ

(photons1/2cm-1sec-1/2)

Y0 (O

)

(Am

ps x

10-9

) 2

θ (CH3OH) = 0 ML

Effect of CHEffect of CH33OH (a) Hole Traps on OH (a) Hole Traps on Initial OInitial O22 PhotodesorptionPhotodesorption YieldYield

The critical point (where slope changes) is reflective of the filling of the active trap sites.

CH3OH causes break point to move to higher Fhυ.

The critical point (where slope changes) is reflective of the filling of the active trap sites.

CH3OH causes break point to move to higher Fhυ.

Summary

• Charge carrier dynamics monitored by quantitatively studying a simple model photoreaction – O2 photodesorption.

• F rate law found – caused by second-order e-h recombination kinetics.

• Artificial enhancement of hole trapping by adding CH3OH – a hole trap molecule.

h υ½

QMSHg Lamp

and Filters

Photodiode Detector

QuartzReflector

Fhν

0.096 Fhν

Ta

TiO2(110)

18O2 Photodesorption from TiO2(110)

0 10 20 30 40

0

2

4

6

8

10

12

14

Y

(A x

10-9

)Time (s)

Fhν = 4.08 x 1014 photons cm-2s-1

Ehν = 3.4 +/- 0.05 eV

T = 110 K

O2

hυ on

18O2

YO (in 0.1s)2

o

Kinetics of Hole TrappingKinetics of Hole Trapping--As Studied by OAs Studied by O22 PhotodesorptionPhotodesorption

•• Holes are either partially filled or completely filled by theHoles are either partially filled or completely filled by thetime the first point is measured [in time the first point is measured [in ΔΔt (sampling)].t (sampling)].

•• ΔΔt (sampling) = 0.10 secondst (sampling) = 0.10 seconds

•• Therefore, at Therefore, at FFhhνν((critcrit.):.):

•• 0.10 s x 0.10 s x FFhhνν(critical) = # photons needed to saturate holes in (critical) = # photons needed to saturate holes in photon penetration depth.photon penetration depth.

•• Photon penetration depth = ~100Photon penetration depth = ~100ÅÅ x 10x 10--88cm cm ÅÅ--11 = 10= 10--66 cmcm

•• Therefore, density of traps = ~ Therefore, density of traps = ~ FFhhνν((critcrit) x ) x ΔΔtt / 10/ 10--66 cm = ~ 3 x cm = ~ 3 x 10101818 cmcm--33

ConclusionConclusion-- OO22 PhotodesorptionPhotodesorption Detector of Detector of Hole Trapping PhenomenonHole Trapping Phenomenon

•• First highlyFirst highly--controlled study of charge carrier controlled study of charge carrier trapping effect on TiOtrapping effect on TiO22 single crystal.single crystal.

•• Hole trapping strongly inhibits surface Hole trapping strongly inhibits surface photoreaction.photoreaction.

•• Hole trap density estimated to be ~3x10Hole trap density estimated to be ~3x101818cmcm--33..

•• This hole trap density corresponds to ~3x10This hole trap density corresponds to ~3x10--55

fraction of atomic sites in the crystal bulk.fraction of atomic sites in the crystal bulk.

•• So far you have seen only kinetic So far you have seen only kinetic

measurements of the initial rate of Omeasurements of the initial rate of O22(a) (a)

photodesorptionphotodesorption –– Y at constant Y at constant θθ

•• Studies of the rate of Studies of the rate of photodesorptionphotodesorption asas

a function of time and the depletion of a function of time and the depletion of

OO22(a) have yielded exciting new results.(a) have yielded exciting new results.

OO22

OO

OO22

O2 Photodesorption from TiO2(110):k1 Fhν

hν + TiO2 → e + h k2

h + T → T+ (hole capture by a hole trap)k3

e + h → heat (on recombination sites) k4

h + O2-(a) → O2(g) ↑

Yield of photodesorbing oxygen scales proportionally with the square root of the incident light intensity:

][][

2

2 2/12/1

32

14 Oh

O Fkk

kkdt

dθ

θυ⎟⎟

⎠

⎞⎜⎜⎝

⎛+

−=

Photoexcitation Behavior Above Fhv (crit.) in Kinetic Regime B

h + T T+k2

• Upon saturation of hole traps above Fhv (crit.)

k2 0

][][

2

2 2/12/1

32

14 Oh

O Fkk

kk

dtd

θθ

υ⎟⎟⎠

⎞⎜⎜⎝

⎛+

−=

[ ]∫ ∫ ⎟⎟

⎠

⎞⎜⎜⎝

⎛−= 2

121

3

14

2ln

υ

θh

O Fkk

kdt

d

[ ] tkk

kF O

h

⋅⎟⎟⎠

⎞⎜⎜⎝

⎛−=⋅

2/1

3

142/1 2

ln1 θυ

1.0 0.8 0.6 0.4 0.2

0

2

4

6

8

10

12

k 4 (k1/k

3)1/2 (x

10-9) (

arb.

uni

ts)

Relative O2 Coverage [ θ O ]2

Time (s)500 250

Range of Experimentally Measured Range of Experimentally Measured PhotodesorptionPhotodesorptionRate ConstantsRate Constants

k4 varies by a factor of ~100 over 250 seconds and is

shown to be time dependent, thus exhibiting fractal

character.

Time-dependent rate constants are described by a Fractal Kinetic Process

Rate Coefficient

“Percolation Cluster” R. Kopelmann, Science 241 (1988) 1620.

R. Kopelman in “Fractal Approach to Heterogeneous Chemistry” (1989) 295.

R. Kopelman, J. Stat. Phys. 42 (1986) 185.

101 ≤≤= − hwheretkk h

instantaneous rate coefficient

fractal term

h = 0 – infinite mixing

h = ½ - one dimensional transport

htkk −⋅= 1

.lnln constthk +−=

ln k versus ln t should give a straight line with slope = -h

0 5

-27

-24

-21

-18

h = 0.33

h = 0.48

ln [k

4 (k1/k

3)1/2 ] (

arb.

uni

ts)

ln(time, (s))

Fractal Kinetic Behavior of the Photodesorption of O2 from TiO2(110)

k4 = k04 t

-0.51 +/- 0.03

h = 0.54

(2-D Fractal)

(1-D Fractal)

-

--

“k”4 = k4t -0.51±0.03

A tentative new model explaining the A tentative new model explaining the 11--D fractal OD fractal O22 desorptiondesorption kinetics is kinetics is proposed.proposed.

T. L. Thompson and J. T. Yates, Jr., T. L. Thompson and J. T. Yates, Jr., ““Control of Control of a Surface Photochemical Process by Fractal a Surface Photochemical Process by Fractal Electron Transport Across the Surface: OElectron Transport Across the Surface: O22PhotodesorptionPhotodesorption from TiOfrom TiO22(110),(110),”” J. Phys. J. Phys. Chem. B Chem. B 110110, 7431 (2006)., 7431 (2006).

defect state

h+

e-

Ef

Filled electronic state at defect site

e- e- e-

h+

O2-

TiO2

e- e-

Fractal Electron Transport via O-Vacancy Sites

• “Charge Cloud” associated with vacancy - Vijay and Metiu, J. Chem. Phys. 118 (2003) 6536.

• “Charge Cloud” causes enhanced electron transport across the crystal surface.

• Reduced efficiency of CT between photoproduced holes and adsorbed O2-.

Besen2.movBesen2.mov

0

30

60

90

120

150

180

210

240

270

300

330

<110>

40

30

20

10

<001>

_

O. Byl and J. T. Yates, Jr. 2006

Anisotropy in the Bulk Resistivity of TiO2(110) Crystal –Four-point Probe Measurements at 300 K

Summary and Conclusions:

•Charge clouds surrounding surface oxygen vacancies combine to form a percolation network for electron transport across the surface causing a decrease in the efficiency of photodesorption of O2.

•A kinetic analysis of the O2 photodesorption process reveals that the rate coefficient for photodesorptionchanges by a factor of > 100 over ~ 250 seconds.

•Electron transport across the surface is likely to be one-dimensional, based on the fractal analysis. Significant anisotropy in conduction through TiO2 has been observed in bulk measurements.

Studies on TiO2 Powder

• ESR

• IR

CB

- - -- - - - - -

hνshallow trapsT = 90K

T. Berger, M. T. Berger, M. SterrerSterrer, O. , O. DiwaldDiwald, E. , E. KnKnöözingerzinger, , D. D. PanayotovPanayotov, T. L. Thompson and J. T. Yates, Jr., T. L. Thompson and J. T. Yates, Jr.

J. Phys. Chem. B J. Phys. Chem. B 109109, 6061 (2005), 6061 (2005)

VB+ + + +

Ti3+

O-

Correlation Between Correlation Between PhotogeneratedPhotogeneratedElectron and Hole Production at 77K in TiOElectron and Hole Production at 77K in TiO22

PowderPowder

EPR Signature for Ti3+ and O-

OO2 2 (a)(a) as an Electron Scavenger:as an Electron Scavenger:Production of OProduction of O22

-- (ads.)(ads.)

TiO2 + O2(a)

lattice O- (hole)+

scavenged e (O2-(a))

hν

IR Spectroscopy Can Be Used IR Spectroscopy Can Be Used To See Electrons Excited Into To See Electrons Excited Into

Conduction BandConduction Band

UV-Induced Increase of IR Background – CB Trapping

CB

VB

2500 2000 1500 1000

IR A

bsor

banc

e

Wavenumber, cm-1

+

a. b.

++

hν hν hν

T. Berger, M. T. Berger, M. SterrerSterrer, O. , O. DiwaldDiwald, E. , E. KnKnöözingerzinger, , D. D. PanayotovPanayotov, T. L. Thompson and J. T. Yates, Jr., T. L. Thompson and J. T. Yates, Jr.

J. Phys. Chem. B J. Phys. Chem. B 109109, 6061 (2005, 6061 (2005))

IR

UV

UV on

UV off

Conduction Band Electrons Transfer to Adsorbed Organic Molecules

CB

VB

ClC2H4SC2H5

e-

-0.2 0.0 0.2 0.4 0.6 0.8 1.0 1.20.6

0.7

0.8

0.9

1.0

1.1

Bac

kgro

und

Abs

orba

nce

at ν

= 2

146

cm-1

Admitted Gas Pressure (Torr)

TiO2-SiO2

T = 200 K

Adsorbed 2-CEES

Adsorbed DES

SC2H5

C2H5

Preferential Electron Transfer: TiO2 CB Electrons to Electrophilic Molecules

Preferential Electron Transfer: TiO2 CB Electrons to Electrophilic Molecules

Xe-

CB

VB

ClC2H4SC2H5

e-

CB

VB

• CB electron transfer to electrophilic atom in adsorbed molecules

SUMMARY

• Oxygen Vacancy Defects – Made by Heating Above 700 K

• O2 Adsorbs on Vacancies

• Hole Traps Retard Photochemistry of O2

• ESR and IR Effective Tools for Detection of Holes [O-1] and CB Electrons in TiO2 Powders

• CB Electrons Populated by UV at 130 K as Seen by IR Background Shift

• CB Electrons Depopulated by Electrophilic Organic Adsorbates – need electronegative atom

A Recent Development- Explanation of UV-Induced TiO2 Hydrophilicity

Rong Wang, Kazuhito Hashimoto, Akira Fujishima, Makota Chikuni, Eiichi Kojima,

Atsushi Kitamura, Mitsuhide Shimohigoshi, Toshiya Watanabe

Nature, 388 (1997) 870-873.

“Light-Induced Amphiphilic Surfaces”

Taken from: Fujishima, Hashimoto, Watanabe, “TiO2-Photocatalysis, Fundamentals and Applications”, BKC Inc. Tokyo, 1999

UV

Taken from: Hata, Kai, Yamanaka, Oosaki, Hirota, Yamazaki, JSAE Review 21, 97-102, 2000

60% of Toyota automobiles already use this technology today.

UV

TiO2 - UV-induced Hydrophilicity - ApplicationsAnti-fogging

What causes TiO2 to become hydrophilic in UV?3 Models:

•Photocatalytic removal of organic compounds ?UV

hydrophilic TiO2hydrophobic TiO2

•UV induced oxygen vacancies ?Nakajima, Koizumi, Watanabe, Hashimoto, J. Photochem. Photobiol. A 146, 129-132, 2001hydrophilic

OH

Ti Ti

OH+

hν

•UV-Induced bond breaking?

UV

Fujishima, Rao, Tryk, J. Photochem. Photobio C 1, 1 2000.

Sakai, Fujishima, Watanabe, Hashimoto, J. Phys. Chem. B 103, 2188, 1999.

Problems with Existing Understanding of the UV-Induced Hydrophilicity Phenomena on TiO2

All current contact angle measurements have been

made in the ambient atmosphere on surfaces which

are not atomically clean. Hydrocarbon (and other

organic contamination) effects are uncontrolled.

Problem needs a clean surface-ultrahigh vacuum

approach and atomic resolution of surface atoms.

STM Investigation of Vacancy Formation by STM Investigation of Vacancy Formation by Intense UV Irradiation Intense UV Irradiation –– TiOTiO22(110)(110)--(1x1)(1x1)

2.0nm

UV Exposure:

Sergey Mezhenny, Peter Maksymovych, Tracy L. Thompson, Oliver Diwald, Dirk Stahl and Scott D. Walck, Chemical Physics Letters, 369 (2003) 152-158.

2.0nm O-vacancy

Ti row

No Irradiation 5.4x1021 photons/cm2 (hν > 3.0eV)

Q ≤ 10-23.5±0.2 cm2

Implies that ~108 photons/site do not produce O vacancy defects!

LiquidLiquid--Solid Contact Angle MeasurementsSolid Contact Angle Measurements

γSV

γLV

γSLθ

surface

droplet

γSV

γLV

γSLθ

surface

droplet

A γSV increases, θ decreases

This technology allows study of contact angle This technology allows study of contact angle for pure Hfor pure H22O under conditions of:O under conditions of:

-- well controlled initial surface cleanlinesswell controlled initial surface cleanliness

-- well controlled atmospherewell controlled atmosphere

-- well controlled photon fluxwell controlled photon flux

Typical H2O Contact Angle ShowingSudden Onset of Wetting of TiO2(110)

P = 1 atm; hexane = 120 ppmO2

a

b

c

0 s

154 s

155 s

0 50 100 150 200 2500

10

20

30

40

T = 297-302 KP = 1 atm

Fhν= 1.1x1017 photons cm-2s-1 (2.1-4.4 eV)

Phν= 0.049 W cm-2

Region of uncertainty

O2

360 ppmhexane

120 ppmhexane

0 ppmhexane

Hexane Vapor Effect on the UV-Induced Wetting of TiO2(110)

Time (s)

θ C ,

Con

tact

Ang

le (d

egre

es)

hν

t100 t200 t3000

300 ppm C6H14

200 ppm C6H14

TiO2

O2 and C6H14

100 ppm C6H14

Schematic Origin of Wetting Delay Period

Hex

ane

Cov

erag

e

Induction Periods0

hν

t100 t200 t3000

300 ppm C6H14

200 ppm C6H14

TiO2

O2 and C6H14

100 ppm C6H14

Schematic Origin of Wetting Delay Period

Hex

ane

Cov

erag

e

Induction Periods0

0 100 200 300 400 5000

100

200

300

400

+2σ error

-2σ error

R = 0.906

Hexane Effect on the Wetting Delayof UV-Induced Wetting of TiO2(110)

Wet

ting

Del

ay (s

)

Hexane Concentration (ppm)4 data points

P = 1 atmT = 297-302 KFhν

= 1.1x1017 photons cm-2s-1 (2.1-4.4 eV)

Phν= 0.049 W cm-2

O2

ConclusionsConclusions

Hydrophilicity model involving UV model involving UV

production of Oproduction of O--vacancy defect sites is vacancy defect sites is

unlikely based on STM results.unlikely based on STM results.

HydrophilicityHydrophilicity model involving hydrocarbon model involving hydrocarbon

photooxidationphotooxidation to produce clean to produce clean wettablewettable

TiOTiO22 is likely to be true. is likely to be true.

-- Induction period scales with Induction period scales with ppmppm

concentration of hexane in Oconcentration of hexane in O22..

What Have We Learned?What Have We Learned?

Photochemistry on semiconductors occurs by e-hpair production in the substrate, accompanied by charge transfer to adsorbed species

Defect sites are important- On surface for adsorption of molecules- In bulk-promoting charge-carrier recombination

Second-order e-h recombination can give F dependence of photochemical kinetics

Hole traps reduce photochemical efficiency

Fractal photodesorption kinetics – controlled by anisotropy in crystal electrical conductivity

1/2hυ

EPR – detects e & h states sensitivelye = Ti3+

h = O-

IR detects CB electrons – delocalization gives continuous optical absorption in IR region

Hydrophilicity induced by UV on TiO2 – caused by photooxidation of organic layers in equilibrium with hydrocarbons in atmosphere, causing cleanup of TiO2 surfaces

Don’t believe STM images unless you know you have chemical control

AcknowledgementsAcknowledgements

Tracy L. ThompsonTracy L. Thompson University of PittsburghUniversity of PittsburghPeter Peter MaksymovychMaksymovych University of PittsburghUniversity of PittsburghDimitarDimitar PanayotovPanayotov University of Pittsburgh; Bulgarian University of Pittsburgh; Bulgarian

Academy of SciencesAcademy of SciencesSergey Sergey MezhennyMezhenny University of Pittsburgh University of Pittsburgh →→ University of University of

MarylandMarylandTykhonTykhon ZubkovZubkov University of PittsburghUniversity of Pittsburgh→→ PNNLPNNLDirk StahlDirk Stahl University of PittsburghUniversity of Pittsburgh→→ LeitzLeitzOliver Oliver DiwaldDiwald University of PittsburghUniversity of Pittsburgh→→ T. U. Wien, T. U. Wien,

AustriaAustriaProfessor Erich Professor Erich KnKnöözingerzinger T. U. Wien, AustriaT. U. Wien, AustriaThomas BergerThomas Berger T. U. Wien, AustriaT. U. Wien, AustriaMartin Martin SterrerSterrer T. U. Wien, AustriaT. U. Wien, Austria

This work has been supported by the Army Research Office underThis work has been supported by the Army Research Office underDr. Stephen Lee and DARPA; Also by PPG IndustriesDr. Stephen Lee and DARPA; Also by PPG Industries