Progress & challenges in plasmonenhanced photocatalysis and photovoltaics

Jen Dionne
Jon Scholl, Andrea Baldi, Ashwin Atre, Di Wu, Justin Briggs, Michael Wisser, Aitzol Garcia, Ai Leen Koh, Tim Burke, Alberto Salleo, Mike McGehee
Materials Science & Engineering | Stanford University

Oh the places plasmons go!

Jen Dionne
Jon Scholl, Andrea Baldi, Ashwin Atre, Di Wu, Justin Briggs, Michael Wisser, Aitzol Garcia, Ai Leen Koh, Tim Burke, Alberto Salleo, Mike McGehee
Materials Science & Engineering | Stanford University

Oh, the places plasmons go!

Oh, the plasmons go quantum When particles are small Their spectra shift blue Their peaks are less tall And because theyre sensitive To their surroundings and charge Catalytic sensing is easy On particles small and large. Well coat titania over a metallic core: UV light gets absorbed e-/h+ pairs separate more But solar photons span Wavelengths red, green, and blue For efficient PV Well use upconversion too!
TiO2 Ag 10 nm

uncharged Ag Intensity (a.u.)

50 eadded Wavelength (nm)

Plasmon resonances of conducting nanoparticles

Plasmon resonances of conducting nanoparticles

Electric field

Plasmon resonances of conducting nanoparticles

r<<l

Plasmon resonances of conducting nanoparticles

imag

real

Plasmon resonances of conducting nanoparticles

10 nm

8x10 nm

Ag

Plasmon resonances of conducting nanoparticles

10 nm H2O Ag Ag

Plasmon resonances of conducting nanoparticles

5 nm H2O Ag

Plasmon resonances of conducting nanoparticles

5 nm

+ 50 e H2O Ag

Plasmon resonances of conducting nanoparticles

5 nm

+ 100 e H2O Ag

Plasmon resonances of conducting nanoparticles

5 nm

+ 200 e H2O Ag

Plasmon resonances and energy conversion

Near-field enhancement

Hot electrons
e fS EC EF EV Au Si

Au 50 nm 100 nm
A

Thomann, Brongersma, Nano Lett. (2011)

Knight, Halas, Science (2011); see also Mubeen, Moskovits, Nat. Nano. (2013)

Reaction Sensors
Scattered intensity

Pd

PdH

Pd

D- D+e-Au
Au
Au

Peak wavelength

830,000 e-

50 nm

100 nm

100 nm

Wavelength
Tang, Liu, Dionne, Alivisatos, JACS (2011)

Reaction time (min)

Novo et al., Nature Nanotech 3(10) 2008

Plasmon resonances and energy conversion

0.6 Catalyst activity (mmol CO/gAu)
Al2O3 Fe2O3 MgAl2O4 SiO2 TiO2

0.3

Bulk Au: catalytically inactive

0
0 5 15 20 25 10 Diameter of Au particles (nm)
Norskov, NanoToday, 2007

oxidation 5 m

30

1. Can we detect plasmons from particles in the sub-10nm regime?

2. Can we use these plasmons to monitor photocatalytic reactions in-situ?

3. Can we improve below-bandgap absorption of solar photons for photocatalysis & photovoltaics?

Plasmon catalysis into the single-nm regime

Probing the plasmonic properties of very small particles is challenging: weak optical scattering, particle heterogeneity in ensemble, organic ligands, etc. Ensemble Measurements:
17.8 nm oleylamine

Ag

2.2 nm (x420)

20 nm
Peng, Schatz PNAS (2010)

Plasmon catalysis into the single-nm regime

Probing the plasmonic properties of very small particles is challenging: weak optical scattering, particle heterogeneity in ensemble, organic ligands, etc. Ensemble Measurements:
17.8 nm oleylamine

Single Particle Measurements:

D=
Diameter

Ag D=

2.2 nm (x420)

D=

20 nm
Peng, Schatz PNAS (2010) Lindfors, Sandoghdar, PRL (2004)

Probing very small plasmonic particles: EELS

Monochromated STEM Electron Beam

Sample on Thin Carbon Film

Jon Scholl, Ai Leen Koh

Annular Dark Field Detector

EEL Spectrometer Prism

CCD

Scanning Transmission Electron Microscopy (STEM) EELS has an imaging spatial resolution of ~0.25nm

Individual, organic-ligand-free nanoparticles

Organic-ligand-free synthesis minimizes organic contamination and the influence of ligand surface damping

EELS: Classically-sized Particles

Counts (a.u.) 2

3.5 Energy (eV)

EELS: Classically-sized Particles

Surface resonance

Counts (a.u.) 2

3.5 Energy (eV)

EELS: Classically-sized Particles

Surface resonance

Counts (a.u.) 2

3.5 Energy (eV)

EELS: Classically-sized Particles

Surface resonance Bulk resonance

Counts (a.u.) 2

3.5 Energy (eV)

EELS: Classically-sized Particles

Surface resonance Bulk resonance

Counts (a.u.) 2

3.5 Energy (eV)

EELS: Size-dependent spectral response

Counts (a.u.) 2.5 3.5 Energy (eV)

4.5

5 nm

Scholl, Koh, Dionne, Nature 483 (2012)

EELS: Size-dependent spectral response

20

Particle Diameter (nm) 3.5 Energy (eV) 4.5

15

Counts (a.u.)

10

2.5

0 5 nm

3.2

3.4 3.6 3.8 Peak Energy (eV)

EELS: Size-dependent spectral response

Surface 20
20 Diameter (nm)

Bulk

Particle Diameter (nm)

15

Counts (a.u.)

10

3.7 3.9 4.1 Energy (eV)

2.5

3.5 Energy (eV)

4.5

0 5 nm

3.2

3.4 3.6 3.8 Peak Energy (eV)

Modeling the size-dependence

AvF Classical treatment uses damping term: Bulk R Accounts for peak broadening, predicts a red shift in noble metals Instead, use a quantum approach:

J. Garcia de Abajo, Nature 483 (2012)

Quantum theory matches experiment

Modeling the impact of these discrete electron transitions on particle spectra results in a blue shift comparable to experiments
1 Particle Diameter (nm) 20 15 10 5 3 3.2 3.4 3.6 3.8 Energy (eV) 4 Normalized absorption efficiency 0

Quantum theory matches experiment

Modeling the impact of these discrete electron transitions on particle spectra results in a blue shift comparable to experiments
1 Particle Diameter (nm) 20 15 10 5 3 3.2 3.4 3.6 3.8 Energy (eV) 4 Normalized absorption efficiency 0

Quantum theory matches experiment

Modeling the impact of these discrete electron transitions on particle spectra results in a blue shift comparable to experiments
1 Particle Diameter (nm) 20 15 10 5 3 3.2 3.4 3.6 3.8 Energy (eV) 4 Normalized absorption efficiency 0

Quantum theory matches experiment

Modeling the impact of these discrete electron transitions on particle spectra results in a blue shift comparable to experiments
1 Particle Diameter (nm) 20 15 10 5 3 3.2 3.4 3.6 3.8 Energy (eV) 4 Normalized absorption efficiency 0

Quantum theory matches experiment

Modeling the impact of these discrete electron transitions on particle spectra results in a blue shift comparable to experiments
1 Particle Diameter (nm) 20 15 10 5 3 3.2 3.4 3.6 3.8 Energy (eV) 4 Normalized absorption efficiency 0

Quantum theory matches experiment

Modeling the impact of these discrete electron transitions on particle spectra results in a blue shift comparable to experiments

Analytic
Particle Diameter (nm)
Particle Diameter (nm) 20 15 10 5 3 3.2 3.4 3.6 3.8 Energy (eV) 4 20 15 10 5 3 3.2

Ab-initio

1 Normalized absorption efficiency 0

3.4 3.6 3.8 Energy (eV)

DFT permittivity functions based on He & Zeng, JPCC. 2010 Scholl, Koh, Dionne, Nature 483 (2012)

Plasmon resonances and energy conversion

0.6 Catalyst activity (mmol CO/gAu)
Al2O3 Fe2O3 MgAl2O4 SiO2 TiO2

0.3

Bulk Au: catalytically inactive

0
0 5 15 20 25 10 Diameter of Au particles (nm)
Norskov, NanoToday, 2007

oxidation 5 m

30

1. Can we detect plasmons from particles in the sub-10nm regime? Yes! Localized plasmons become quantum confined around 5 nm Single electron transitions can significantly impact the collective response of a sea of electrons

Plasmon resonances and energy conversion

0.6 Catalyst activity (mmol CO/gAu)
Al2O3 Fe2O3 MgAl2O4 SiO2 TiO2

0.3

Bulk Au: catalytically inactive

0
0 5 15 20 25 10 Diameter of Au particles (nm)
Norskov, NanoToday, 2007

oxidation 5 m

30

1. Can we detect plasmons from particles in the sub-10nm regime?

2. Can we use these plasmons to monitor photocatalytic reactions in-situ?

Case study: water-splitting photocatalysis

Porous TiO2

IrO2nH2O

Youngblood et al., JACS (2009)

Case study: water-splitting photocatalysis

Porous TiO2

O2

H2O

IrO2nH2O

Youngblood et al., JACS (2009)

Can small plasmonic particles help?

Ag

TiO2 C.B.

EF V.B.

EtOH EtOH
Andrea Baldi

1. Synthesis of well-dispersed Ag@TiO2 nanoparticles 2. Characterization of their photocatalytic activity in: Ensemble measurements Single particle measurements

Synthesis of Ag@TiO2 nanoparticles

200 nm

20 nm

Synthesis of Ag@TiO2 nanoparticles

Ag

200 nm

20 nm

Synthesis of Ag@TiO2 nanoparticles

Ag Ag TiO2

200 nm

20 nm

Synthesis of Ag@TiO2 nanoparticles

Ag Ag TiO2

200 nm

20 nm

Ensemble Measurements
UV irradiation of de-aerated Ag@TiO2
Hg(Ne) lamp

254

313 365 405 436 Wavelength (nm)

19 nm

See also: Kamat et al., JACS (2005); ACS Nano (2011)

Ensemble Measurements
UV irradiation of de-aerated Ag@TiO2

Ag

TiO2

EtOH EtOH
19 nm

See also: Kamat et al., JACS (2005); ACS Nano (2011)

Ensemble Measurements
Discharge in O2

Ag

TiO2

O2 O2

See also: Kamat et al., JACS (2005); ACS Nano (2011)

Single Particle Optical Measurements

Single Particle Optical Measurements

A B

Single Particle Optical Measurements

Charge 0 min UV

Single Particle Optical Measurements

Charge 20 min UV

Single Particle Optical Measurements

Charge 40 min UV

Single Particle Optical Measurements

Charge 60 min UV

Single Particle Optical Measurements

Discharge 0 min O2

Single Particle Optical Measurements

Discharge 20 min O2

Single Particle Optical Measurements

Discharge 40 min O2

Single Particle Optical Measurements

Discharge 60 min O2

Single Particle Optical Measurements

Discharge 80 min O2

Single Particle Optical Measurements

Ensemble Particle A Particle B Particle C

Bars denote peak full-width at half maximum On-going: correlate single particle structure with catalytic activity

Plasmon resonances and energy conversion

0.6 Catalyst activity (mmol CO/gAu)
Al2O3 Fe2O3 MgAl2O4 SiO2 TiO2

0.3

Bulk Au: catalytically inactive

0
0 5 15 20 25 10 Diameter of Au particles (nm)
Norskov, NanoToday, 2007

oxidation 5 m

30

1. Can we detect plasmons from particles in the sub-10nm regime?

2. Can we use these plasmons to monitor photocatalytic reactions in-situ?

Preliminary results are promising. It will be exciting to correlate catalyst size and shape with activity.

Plasmon resonances and energy conversion

0.6 Catalyst activity (mmol CO/gAu)
Al2O3 Fe2O3 MgAl2O4 SiO2 TiO2

0.3

Bulk Au: catalytically inactive

0
0 5 15 20 25 10 Diameter of Au particles (nm)
Norskov, NanoToday, 2007

oxidation 5 m

30

1. Can we detect plasmons from particles in the sub-10nm regime?

2. Can we use these plasmons to monitor photocatalytic reactions in-situ?

3. Can we improve below-bandgap absorption of solar photons for photocatalysis & photovoltaics?

Solar upconversion
Incident Power (W m-2)

1000
5%

800
25%

Power Transmitted

600
50%

400 200 0
75%

Solar cell

Power to Cell Power to device 2 3 Bandgap, Eg (eV) 4

100% Absorption >Eg

Photons with an energy less than the bandgap (Eg) of a solar cell are unused Photons with energy just above Eg are poorly absorbed

5 % Ultraviolet

43 % Visible

52 % Infrared

A potential solution: upconversion

Solar cell Upconverter

Solar cell

Insulator
Upconverter

A process where light is emitted with photon energies higher than those of the light generating the excitation Implementation: Solar cell and upconverter are electrically isolated 1) No additional recombination in cell 2) Separate optimization of cell & upconverter
Trupke, et al. J. Appl. Phys. 92 (2002)

Upconverting cells exceed the Shockley-Queisser limit

44 30

With upconverter

% increase in efficiency

300

Cell efficiency (%)

Solar concentration: 103 102 10 1

200

No upconverter

100

1.0

1.5 2.0 Cell bandgap (eV)

2.5

1.0

2.0 Cell bandgap (eV)

3.0

Peak cell efficiency increases from 30% to 44% Ideal cell bandgap blue-shifts from 1.1 eV to 1.8 eV Solar concentration is not necessary
Trupke, et al. J. Appl. Phys. 92 (2002) A. Atre and J. Dionne. J. Appl. Phys. 115 2 (2011)

How does non-ideal upconversion impact a cell?

Cell Bandgap (eV)

Cell Efficiency (%)

cell Eg=1.7 eV

Thermodynamic limit (no UC)

UC Bandwidth (eV)

Effect of UC absorption/recombination efficiency: need high quantum efficiencies

A. Atre and J. Dionne. J. Appl. Phys. 115 2 (2011)

Effect of UC absorption bandwidth: higher efficiencies with higher bandwidths

J. Briggs, A. Atre and J. Dionne. Submitted (2013)

Additional design considerations

6

5
UC FWHM=0.1eV

Ideal Peak Position (eV)

Increase in Jsc (mA/cm2)

4 3 2 1
Bandgap (slope=1)

Cell Efficiency (%)

Emitter

Absorber 2

Absorber 1

0 0

2000 4000 3000 1000 UC absorption peak location (nm)

0 1

1.5 2 2.5 Cell Bandgap (eV)

Solar-spectrum matching: dont want UC energy levels to overlap with AM1.5 absorption lines
T. Burke, M. McGehee

UC absorption peak positions: usually in the near-infrared (811nm and 1200nm for a 1.7eV cell)
J. Briggs, A. Atre and J. Dionne. Submitted (2013)

Two promising upconverting systems, part 1

Bimolecular systems

PdOEP

DPA

Upconverted PL

Intensity (au)

PDOEP absorption

PdOEP

PdOEP+DPA

Wavelength (nm)

Photos by Ashwin Atre; and Diane Wu See also: Singh-Rachford, et al. JACS. 131 (2009)

Two promising upconverting systems, part 2

Lanthanoid-doped nanoparticles

200 nm Upconverted PL

Intensity (au)

Dyesensitized Ln3+ absorption

400

Photos by Diane Wu Crystal structure from Wang, Nature 2010; Spectra from Hummelen, Nature Photonics 6(2012)

Wavelength (nm)

650

900

What are the predicted improvements?

Bimolecular systems
Absolute % increase in cell efficiency Cell Bandgap (eV) Cell Bandgap (eV) Absolute % increase in cell efficiency Cell Bandgap (eV) Cell efficiency (%) Wavelength (nm) Intensity (a.u.)

Lanthanide systems
Cell efficiency (%) Wavelength (nm)
J. Briggs, A. Atre and J. Dionne. Submitted (2013)

Intensity (a.u.)

Cell Bandgap (eV)

Towards Enhanced Upconversion

Conductive upconverting composites
Diane Wu UC Film Composite UC film (as conductive as ITO)

Intensity (a.u.)

Ag NW

4.1x

4.7x 4.3x

5 m
500 550 600 650 Wavelength (nm) 700

Designing a tunable, broadband plasmonic resonator

Increasing wire aspect ratio

100 nm 300 900 600 Wavelength (nm) 1200

Nanowires are promising, but have a strong polarization and angle-dependent response. Further, need UC in near-field of wire.

Absorptance (a.u.)

Upconverting metamaterials in solar cells

Solar cell
Insulator Upconverter Ag shell, upconverterdoped core Intensity (E2) enhancement of ~100x throughout the entire upconverting volume
Atre, Alaeian, Garcia, Dionne. Journal of Optics (Invited Review), January 2012

Upconverting metamaterials in solar cells

Solar cell
Insulator Upconverter Ag shell, upconverterdoped core

Upconverting metamaterials in solar cells

Solar cell
Insulator Upconverter Ag shell, upconverterdoped core

Upconverting metamaterials in solar cells

No nanocrescent

With nanocrescent
Upconverted power

Solar cell

Solar cell

Crescent Orientation
50 nm 50 nm

Nanocrescent array: a tunable index metamaterial

max Out-of-plane magnetic field (Hz)

The refractive index of the upconverting metamaterial can be matched to the solar cell above it.
A. Atre, A. Garcia, H. Alaeian, J. Dionne, Advanced Optical Materials, in press (2013)

min

Nanocrescent fabrication

100 nm

1 m

Nanocrescent fabrication

1 m

100 nm

Nanocrescent characterization
Intensity (a.u.)
Intensity (a.u.)

2.5 2

1.5 1
400 400

Ensemble absorption Single particle scat.

600

e-

800 1200 1600 800 1000 1200 1400 1600 1800 Wavelength (nm) Wavelength (nm) =700 nm

=600 nm

In collaboration with A. Polman, FOMInstitute AMOLF (Amsterdam)

800 nm

900 nm

Coming soon: upconverting crescents

Bimolecular systems Lanthanoid systems

100 nm
Li, Z. Nanotechnology 2008

Upconversion at 445 nm
Synthesis based on Wohnhass, Macromol. Biosci 11, 772 (2011)

50 nm

Summary
Absorption Particle Diameter (nm) 20 15 10 5

1. Can we detect plasmons from particles in the sub-10nm regime? 2. Can we use these plasmons to monitor photocatalytic reactions in-situ? 3. Can we improve below-bandgap absorption of solar photons for photocatalysis & photovoltaics? Some see things as they are and ask why?. Others dream things that never were and ask, why not?. George Bernard Shaw.

3.5 Energy (eV)

1 m

Thanks to our funders: Stanford, AFOSR, NSF, SLAC/SIMES, DOE, GCEP, Intel!

Summary
Absorption Particle Diameter (nm) 20 15 10 5

1. Can we detect plasmons from particles in the sub-10nm regime? 2. Can we use these plasmons to monitor photocatalytic reactions in-situ? 3. Can we improve below-bandgap absorption of solar photons for photocatalysis & photovoltaics? Some see things as they are and ask why?. Others dream things that never were and ask, why not?. George Bernard Shaw.

3.5 Energy (eV)

1 m

Thanks to our funders: Stanford, AFOSR, NSF, SLAC/SIMES, DOE, GCEP, Intel!

Bimolecular upconversion process

S = sensitizer; E = emitter
h1
Energy Requirements:

1S 1S S*
ISC

3S*

TET

1E

1S* 3S* h1 1S*

ISC

1E* 3E*
TTA

1E*
1E 1E

h2

3E* 3E* 3S* 1E

Sensitizer Emitter
TET

1S

1S

The need for efficient upconversion

6

5
Ideal Peak Position (eV)
All FWHM=0.1eV

Increase in Jsc (mA/cm2)

4 3 2 1
Bandgap (slope=1)

Cell Efficiency (%)

Emitter

Absorber 2

Absorber 1

0 0

2000 4000 3000 1000 UC absorption peak location (nm)

0 1

1.5 2 2.5 Cell Bandgap (eV)

Solar-spectrum matching: dont want UC energy levels to overlap with AM1.5 absorption lines
Burke, McGehee. In preparation (2012)

UC absorption peak positions: usually in the near-infrared (811nm and 1200nm for a 1.7eV cell)
Briggs, Dionne. In preparation (2012)

Tackling low UC efficiencies

1. The electronic approach: modify dipole transition moments consider Yb3+/Er3+ all UC transitions are between f levels LaPorte forbidden or spin-forbidden Can we modify the optical transition rates by changing the crystal field?
2F 5/2

Maverick Chea
2H 11/2 4S 3/2 4F 9/2

Michael Wisser

654 nm 542 nm 520 nm

4I 11/2

980 nm
2F

4I 13/2

7/2

Yb3+

Er3+

4I 15/2

Ln3+ upconverters under pressure

Yb3+/Er3+ in hexagonal NaYF4
1.15 GPa 6.97 GPa 12.9 GPa 19.6 GPa 28.3 GPa
100 nm

Ruby sample
500 500 550 550 600 600 650 650

Intensity (a.u.) Intensity (a.u.)

Wavelength (nm)

Wavelength (nm)

Ln3+ upconversion & PL under pressure

Upconversion
537.5

Photoluminescence
540 540
540
542.5
Wavelength (nm)

540 540
Wavelength (nm) Wavelength(nm)

16000 14000 12000 10000 8000 6000 4000

12.9 28.3 14.1 1.15 16000 14000

Intensity (a.u.)

545 545

545 545 545

12000 10000 8000 6000 4000 2000

547.5

550 550

550 550

550

555 552.5

555 555 1.15 1.15 12.9 28.3 12.9 28.3 Pressure (GPa) (GPa) Pressure

555 555 1.15

2000 Pressure (GPa) 0.58 13.5 18.4 25.8 20.2 9.34 2.87 0.58 13.5 18.4 25.8 20.2 9.34 2.87 14.1 1.15Pressure (GPa) Pressure (GPa)

Intensities are affected by pressure: changes to interionic separations and Yb3+/Er3+ energy level resonance

Ln3+ upconverters under pressure

UC Peak Shifts as a Function of Pressure
5 4
12 3 4 5 7

Upconversion
6

PL Peak Shifts as a Function of Pressure

5 4
1 2 3 45

Photoluminescence

Wavelength shift (nm)

3 2 1 0 -1 0

Wavelength shift (nm)

15 20 25

3 2 1 0 -1 0

10

10

15

20

25

Pressure (GPa)

Pressure (GPa)

Peak shifts are also pressure-dependent: peaks spread apart as pressure is increased, suggesting a distortion in the crystal field