Clavero 2014 Nature Photonics

7/21/2019 Clavero 2014 Nature Photonics

1/9

NATURE PHOTONICS| VOL 8 | FEBRUARY 2014 | www.nature.com/naturephotonics 95

The outstanding light-trapping and electromagnetic-field-concentrating properties o surace plasmons open up a widerange o applications in the field o plasmonics1. Localized

surace plasmon resonance (LSPR) can occur in properly designednanostructures in which confined ree electrons oscillate with thesame requency as the incident radiation and eventually enterresonance, giving rise to intense, highly localized electromagneticfields. Consequently, such nanostructures have been proposed asefficient light-trapping components that can be integrated in pho-tovoltaic cells to increase the efficiency o conventional architec-tures considerably2,3. However, recent investigations have shown

that plasmonic nanostructures can also directly convert the col-lected light into electrical energy by generating hot electrons413.Afer light absorption in the nanostructures and LSPR excitation,plasmons can decay, transerring the accumulated energy to elec-trons in the conduction band o the material. Tis process produceshighly energetic electrons, also known as hot electrons, which canescape rom the plasmonic nanostructures and be collected by, orexample, putting the plasmonic nanostructures in contact with asemiconductor, thereby orming a metalsemiconductor Schottkyjunction6. Tis new scheme or solar energy conversion opens upa way to realize photovoltaic and photocatalytic devices whoseperormances may rival, or even exceed, those o conventionaldevices. However, some difficulties and limitations inherent to thenature o this energy conversion process and to the properties o

the materials employed need to be addressed in order to achievelarger efficiencies while keeping abrication costs low.

Plasmonic energy conversionPlasmonic hot-electron generation. Electrons not in thermalequilibrium with the atoms in a material are requently reerred toas hot electrons. Te term hot electrons reers to electrons whosedistributions can be undamentally described by the Fermi unc-tion, but with an elevated effective temperature14. When a materialis illuminated with highly energetic photons (or example, ultra-violet radiation), hot electrons are generated and emitted rom the

Plasmon-induced hot-electron generation at

nanoparticle/metal-oxide interfaces forphotovoltaic and photocatalytic devicesCsar Clavero

Finding higher efficiency schemes for electronhole separation is of paramount importance for realizing more efficient conver-sion of solar energy in photovoltaic and photocatalytic devices. Plasmonic energy conversion has been proposed as a promisingalternative to conventional electronhole separation in semiconductor devices. This emerging method is based on the genera-tion of hot electrons in plasmonic nanostructures through electromagnetic decay of surface plasmons. Here, the fundamentalsof hot-electron generation, injection and regeneration are reviewed, with special attention paid to recent progress towardsphotovoltaic devices. This new energy-conversion method potentially offers high conversion efficiencies, while keeping fabrica-tion costs low. However, several considerations regarding the materials, architectures and fabrication methods used need to be

carefully evaluated to advance this field.

material via the photoelectric effect 15. Electrons whose energiesexceed the work unction o the material are emitted, producinga photocurrent. However, as the solar spectrum is mostly com-posed o less energetic visible and near-inrared photons, the useo this effect is impractical or photovoltaic devices. Hot electronscan also be generated by exothermic chemical processes16,17, suchas those that occur in dye-sensitized solar cells18. In these cells, adye molecule anchored to a semiconductor absorbs incoming lightand transers energetic charge carriers to the semiconductor 19.

A major breakthrough in this field was the recent discovery ohot-electron generation in plasmonic nanostructures. Following

light absorption and LSPR excitation in these nanostructures,electromagnetic decay takes place on a emtosecond timescale,either radiatively through re-emitted photons20 (Fig. 1a, lef) ornon-radiatively by transerring the energy to hot electrons12,2123(Fig. 1a, right). In the non-radiative process, surace plasmons firstdecay into single-electron excited states. Tis might be ollowed byphotoemission i the electron energy exceeds the work unction othe material24. wo-photon processes have also been described23.Non-radiative decay in noble-metal nanostructures can take placethrough intraband excitations within the conduction band orthrough interband excitations caused by transitions between otherbands (or example, dbands) and the conduction band. However,the d band energy levels respectively lie 2.4 eV and 4 eV below theFermi energy levels or Au and Ag, making interband excitations

considerably more unlikely than intraband excitations25,26

. Teremaining photoexcited electrons relax through electronelec-tron and electronphonon collisions, and are ultimately convertedinto heat24.

Figure 1b depicts the parabolic density o states (DOS) in theconduction band o a plasmonic nanostructure with a Fermienergy EF,M as a unction o energy. Afer non-radiative suraceplasmon decay, electrons rom occupied energy levels are excitedabove the Fermi energy. For example, surace plasmons in Au andAg noble-metal nanostructures can transer energies betweenapproximately 1 eV and 4 eV to hot electrons; this energy depends

Plasma Applications Group, Environmental Energy Technologies Division, Lawrence Berkeley National Laboratory, 1 Cyclotron Road, Berkeley, California

94720, USA. e-mail: [email protected]

REVIEW ARTICLEPUBLISHED ONLINE: 30 JANUARY 2014 | DOI: 10.1038/NPHOTON.2013.238


2/9

96 NATURE PHOTONICS| VOL 8 | FEBRUARY 2014 | www.nature.com/naturephotonics

on the carrier concentration and the size and shape o the nano-structures9,20,27. An efficient mechanism or capturing such hotelectrons is to orm a Schottky barrier with an appropriate semi-conductor. Figure 1c shows a Schottky barrier between a plas-monic nanostructure and an n-type semiconductor, such as iO2.iO2is a good electron-accepting metal oxide because o the highDOS in its conduction band; it thus permits ast electron injec-tion. Hot electrons with energies higher than the Schottky barrierenergy SBcan be injected into the semiconductor with an emis-sion efficiency dependent on their energy25. In addition, tunnellingacross the barrier can take place with a much lower probability28.Te energy needed or hot electrons to overcome the energy bar-rier in this system is considerably smaller than the bandgap o the

semiconductor Eg(res 7,28). Afer injection o hot electrons intothe neighbouring semiconductor, the plasmonic nanostructuresare lef positively charged because o electronic depletion. Anelectron-donor solution or a hole-transporting material (HM) isrequired to be in contact with the nanostructures to transport thegenerated holes to the counter electrode, keeping the charge bal-ance and sustaining an electric current.

As this Review will show, plasmonic devices are not affected bythe thermodynamic actors that limit the efficiencies o conventionalsemiconductor-based devices, and they thus open a new horizon opossibilities in the field o solar energy conversion. Te size, shapeand composition o the plasmonic nanostructures can be adaptedto obtain broad absorption across the whole solar spectrum. Tehigh absorption cross-section o plasmonic nanostructures allows

the thickness o the active zone to be reduced while maintaininga high light-trapping efficiency. In addition, recent studies haveshown that their efficiency rises slightly with increasing tempera-ture because o a higher probability o hot-electron injection29.

First steps towards plasmonic energy conversion. Early stud-ies published in 19964 showed hints o surace-plasmon-inducedcharge separation in noble-metal nanoparticles in contact withiO2. Zhao et al.4 were the first to report an anodic photocur-rent generated by visible-light illumination o a iO2 electrodewith a iO2overlayer containing gold or silver nanoparticles. Tiswas a striking result because the wide bandgap (3.3eV) o iO2means that conventional iO2 electrodes generate photocurrentsonly when illuminated with ultraviolet light. Te origin o thesephotocurrents was unclear at the time; Zhao et al. suggested the

excitation o LSPR in the nanoparticles as a possible mechanism.Tis research field remained dormant or some years until 2003,when multicolour photochromism (that is, areversible change incolour on illumination) in Ag nanoparticles dispersed in iO2wasreported30. Tis phenomenon was ascribed to LSPR-induced chargeseparation and oxidation o the Ag nanoparticles30,31(Fig. 2a). Teinverse process, namelyreversible injection o electrons rom iO2into Ag in coreshell AgiO2 nanoparticles, was also demon-strated by Hirakawa and Kamat 32.

Shortly afer, several studies5,6,33,34demonstrated the three stepsrequired to generate a photoinduced closed-circuit current hot-electron generation, injection and regeneration. Systems consist-ing o Au or Ag nanoparticles absorbed onto nanoporous iO2

(which, in some cases, was in contact with an electron-donor solu-tion) were investigated (Fig. 2b). ian and atsuma5,6 observedhot-electron injection into the iO2matrix on excitation o LSPRin the nanoparticles by visible light illumination. Subsequently,compensative electrons rom a donor solution in contact with theplasmonic nanoparticles were injected into the nanoparticles, bal-ancing their electronic deficit and ultimately creating an electriccircuit. Te measured action spectra, which show the incidentphoton-to-electron conversion efficiency (IPCE) as a unction owavelength, exactly reproduced the extinction spectra o the Aunanoparticles, clearly demonstrating that LSPR plays a key role inpromoting hot-electron generation (Fig. 2b).

Following these pioneering studies, very many studies haveinvestigated plasmonic hot-electron generation with application

to both photovoltaic and catalytic devices. Most o these stud-ies investigated Au or Ag nanoparticles in contact with iO 23345

(Fig. 2c), but more recently multiple combinations o materialsand architectures have been proposed. Some examples are mul-tilayer assemblies o Au nanoparticles and iO2 nanosheets

46, Ptnanoparticles on iO2 thin films

47, Ag-decorated iO2 nanotubearrays48,49 (Fig. 2d), ZnO nanorods decorated with Au nanopar-ticles50,51, hierarchical AuZnO flower-rod heterostructures52(Fig. 2e), Au nanoparticles in contact with CeO2

53,54, Au nano-prisms on WO3

55, AgBr decorated with Ag nanoparticles and dis-persed in Al2O3

56, AgAgI supported on mesoporous alumina57,AgCl particles decorated with Ag nanoparticles58, coreshell SiO2iO2 nanoparticles decorated with Au nanoparticles

59, MiO2coreshell nanocomposites (where M = Au, Pd, Pt) 60,61 (Fig. 2),Au nanorods on Si7and AualuminaAu multilayers8.

ba

Radiative decay Non-radiative decay

DOS

c

d-band

EF,M EF,M

E

M

SB

EF,ME

F,S Eg

Ec

Ev

S

Metal Semiconductor

e

e

ee

e

h+

h+

Figure 1 |Surface-plasmon decay, hot-electron generation and injection. a, Localized surface plasmons can decay radiatively via re-emitted photons

or non-radiatively via excitation of hot electrons. In noble-metal nanostructures, non-radiative decay can occur through intraband excitations within

the conduction band or through interband excitations resulting from transitions between other bands (for example, dbands) and the conduction band.

b, Plasmonic energy conversion: electrons from occupied energy levels are excited above the Fermi energy. c, Hot electrons can be injected into a

semiconductor by forming a Schottky barrier with the plasmonic nanostructure. Hot electrons with energies high enough to overcome the Schottky barrier

SB= M Sare injected into the conduction band Ecof the neighbouring semiconductor, where Mis the work function of the metal andSis the electronaffinity of the semiconductor.

REVIEW ARTICLE NATURE PHOTONICS DOI: 10.1038/NPHOTON.2013.238


3/9


Timescales of hot-electron generation, injection and regenera-tion. Fast-electron injection in the neighbouring semiconductorbeore carrier recombination is a key actor or increasing theefficiency o the energy conversion process. Several studies haveocused on the timescales o the hot-electron generation, injec-tion and regeneration mechanisms. Tey exploit the increased vis-ible and inrared absorbance o iO2when electrons are injected

into the conduction band6,6264

. Furube and Dus group11,62,65

haveused ultraast visible-pump/inrared-probe emtosecond transientabsorption spectroscopy to characterize the charge transer kinet-ics. Laser pulses (duration, 150 s; wavelength, 550 nm) were usedto excite LSPR in Au nanoparticles deposited on iO 2 (Fig. 3a),while the transient absorption o iO2at 3,500 nm was simulta-neously monitored. Te group ound that hot-electron genera-tion and injection completed within 50 s. Because relaxation oelectrons with a non-Fermi distribution in Au nanoparticles wasound to occur through electronelectron (


4/9


efficiency obtained with these plasmonic structures increases withincreasing temperature, in contrast to the general trend o semi-

conductor-based solar cells (Fig. 3b, bottom). Plasmonic energyconversion could thus solve the overheating problem that afflictsconventional photovoltaic cells69.

Several works have provided insights into the localization o hot-electron generation in plasmonic nanostructures. Kazuma et al.70,71showed that sites in Ag nanorods on iO2exposed to higher elec-tromagnetic fields generate more hot electrons. Afer LSPR exci-tation and hot-electron generation, the oxidized Ag+ ions diffusein the water layer adsorbed on the iO2substrate, and eventuallyrecombine with electrons rom iO2, leading to the ormation osatellite redeposited islands (Fig. 3c, top). Tis effect was observedin rods with different aspect ratios that sustained different multi-pole plasmon modes (m is 14), as shown in the atomic orcemicroscopy images shown in the bottom portion o Fig. 3c. Islands

appeared in locations where the LSPR electromagnetic field wasmore intense, confirming the hypothesis that the charge generationprocess is induced and/or promoted by the intense electromagneticfields in the plasmonic nanostructures. Optimizing the design oplasmonic nanostructures so as to maximize the electromagneticfields will increase hot-electron generation.

Te geometry and the locations o the plasmonic nanostruc-tures relative to the neighbouring semiconductor material arealso very important. Knight et al.12showed that the photocurrentgenerated by an active plasmonic element can be significantlyenhanced by embedding it in the neighbouring semiconductor,as this permits more efficient transer o hot electrons. Also, asshown below, direct contact between the donor solution or HMand the semiconductor material should be avoided to prevent det-rimental carrier recombination.

Towards solid-state devices. Electron-donor solutions and HMsplay a key role in regenerating the carriers afer hot-electron injec-

tion. ian and atsuma6

compared the perormances o seven di-erent donor solutions: [Fe(CN)6]4, I, Fe2+, errocenecarboxylic

acid, Br, 1,1-errocenedicarboxylic acid and Cl. Tey reportedan IPCE o 12% at around 560 nm when Fe2+was used as an elec-tron donor. Tey obtained an extremely high IPCE o 26% on theaddition o 4-nitrobenzoic acid, which they attributed to the acidbeing adsorbed on exposed iO2and blocking the recombinationo electrons. Subsequent studies63,64 associated the higher effi-ciency o Fe2+to aster carrier regeneration rates.

Carrier regeneration using liquid electrolytes is advantageousor photocatalysis applications as the liquid electrolyte acilitatesthe flow o the processing chemicals past the plasmonic nano-structures. In contrast, liquid cells are impractical or photovoltaicapplications because o their lack o stability and problems with

leaking and evaporation. In recent years, several solid-state plas-monic solar cell structures have been proposed, bringing practicalapplication one step closer. Initial results or solid-state photovol-taic cells with organic and inorganic HMs were unsatisactory;the efficiencies obtained were very low, being around our orderso magnitude lower than those o liquid cells with an electrolytecontaining a redox couple. Yu et al.72 tested different solid-statecell architectures in which Au nanoparticles are in contact withiO2and various HMs, including poly(N-vinylcarbazole) (PVK),N,N-bis(3-methylphenyl)-N,N-diphenyl-[1,1-biphenyl]-4,4-diamine (PD), CuI and CuSCN. Tey obtained a very lowmaximum IPCE o 0.0024% using PVK, which they attributed tocarrier recombination caused by contact between the HM andiO2. Tey also considered possible degradation o the HM lay-ers. Surprisingly, higher efficiencies were obtained in subsequent

AR: 1.6 3.4 5.7 7.2

m= 1 m= 1 m= 1 m= 1

0.4 0.3 0.2 0.1 0.0 0.1 0.2 0.3

333.15 K323.15 K313.15 K303.15 K293.15 K283.15 K

0

20

40

60

80

100

Dissolution Redeposition

N3/TiO2

Au/TiO2

Au/ZrO2

Absorbance(103)

TiO2

a Hot-electroninjection < 50 fs

e-

e

Photocurrent(nA)

Potential (V vs. SCE)

e Ag e

h

Ag+

TiO2

Ag satelliteNPs

b c

Silicon

Regeneration

< 20 ns

e

e

0

2

4

6

8

10

//

//

IPCE(%)

450

0

4

8

12

20

24

600 800 1,000 1,200Wavelength (nm)

Delay time (ps)

0.5 0.0 0.5 1.0 10 100 1,0001.5

Figure 3 |Hot-electron regeneration and shape effects. a, Ultrafast visible-pump/infrared-probe femtosecond transient absorption spectroscopy shows

the timescale of hot-electron injection in the TiO2conduction band62. b, SEM images for nanopatterned Au nanorods. The dimensions of the Au nanorods

are 100 nm 240 nm 40 nm. The action spectra corresponding to non-polarized light (black), polarized light along the minor (red) and major (blue) axis

directions of the nanorods reproduce the extinction spectra. Also, a progressive increase in the photocurrent is observed with increasing temperature29. c,

Oxidation of Ag nanorods and redeposition of small Ag nanoparticles takes place as a result of plasmon-induced charge separation. After LSPR excitation

and hot-electron generation, the oxidized Ag+ions diffuse in the water layer adsorbed on the TiO2substrate and eventually recombine with electrons from

TiO2, forming satellite redeposited islands. The atomic force microscopy images show Ag nanorods with various aspect ratios (ARs) that have undergone

excitation of multipole plasmon modes from m= 1 to 4 after irradiation with 800 nm. The rods experience changes in length depending on the incident

wavelength and mode excited. Horizontal scale bars are 50 nm (ref. 71). Figure reproduced with permission from: a, ref. 62, 2007 ACS; b, ref. 29, 2010

ACS; c, ref. 71, 2012 ACS.



5/9


studies that did not include any specific electrolyte or HM layer.akahashi and atsuma73,74 showed that devices consisting o Auand Ag nanoparticles on iO2 and in contact with indiumtinoxide (IO), where electrons are simply injected into the nano-particles rom the IO film, can achieve IPCEs o around 0.4%and 0.6% or Au and Ag nanoparticles, respectively. ian et al.64recently produced solid-state solar cells with considerably higherefficiencies (an IPCE as high as 6%) by using a similar architectureconsisting o Au nanoparticles on iO2 and polyethylene oxide

filled with iO2nanoparticles as the HM, which contains opti-mized redox couples I/I3

. Recently, Reineck et al.75used Spiro-OMeAD as the HM film in solid-state cells with sel-assembledAu and Ag nanoparticles, obtaining IPCEs o 4.9 and 3.8%, respec-tively. Although recent advances in this field are encouraging, ur-ther research is needed to obtain HMs and architectures thatprevent carrier recombination occurring through direct contactwith the semiconducting material.

Other applications and approaches. Te injection o hot electronsrom plasmonic nanoparticles into the semiconductor matrix onillumination significantly changes the conductivity o the system.Several studies have explored the possibility o using this principle toobtain plasmo-electronic devices in which the conductivity is con-

trolled by varying the intensity o the visible light illumination. Inrecent studies76,77, this effect has been observed in nanoparticles sta-bilized with different sel-assembled monolayers, which increasedor reduced the electrical conductivity o the material during reso-nance, depending on whether they contained neutral or chargedligands, respectively. Mubeen et al.78proposed devices consisting oAu nanoparticles and iO2multilayers in which hot-electron tun-nelling produced an over 1,000-old increase in the conductance onillumination by 600 nm light. Son et al.79reported surace-plasmon-enhanced photoconductance in iO2 nanofibres loaded with Aunanoparticles. Several other works have ocused on plasmonic pho-toinduced currents in nanodiode structures. Lee et al.80observed anenhanced photocurrent as a result o plasmon-assisted generationo hot electrons on nanodiodes consisting o gold islands on iO 2;

the photocurrent can be urther enhanced by incorporating dyemolecules81. Knight et al.7observed hot-electron flows induced byinrared irradiation o Au/Si nanoantennas. Wadell et al.82investi-gated AuSiO2Pd nanoantennas. Also, manipulation o plasmonicresonance by changing the electronic density has been proposed orapplications such as smart windows and displays83.

So ar, only a ew works have explored plasmonic charge sepa-ration using propagating surace plasmon polaritons (SPPs). Teuse o SPPs is interesting as they cannot decay directly into pho-tons unless surace roughness is present 24; thus, non-radiativedecay is the only mechanism possible in flat films. Wang et al.8proposed a new architecture based on SPP excitation in a metalinsulatormetal structure illuminated under the prism couplerconfiguration. Owing to the high localization o SPPs on the top

metal thin film, more hot electrons are transmitted rom the topelectrode to the bottom electrode than in the opposite direction,leading to the detection o a net photocurrent. Wang et al.scal-culations suggest that this kind o architecture could achieve effi-ciencies o up to 4.3% or 640-nm irradiation o Ag films, and oup to 3.5% or 780-nm irradiation o Au films. Nevertheless, inpractice, they were only able to measure efficiencies o around2.7%, as a result o surace recombination at the metalinsulatorinterace. Along the same lines, Pradhan et al.84investigated semi-conductorinsulatorsemiconductor heterojunctions consistingo Al:ZnO/SiO2/Si.

New directions in plasmonic energy conversionPlasmonic hot-electron generation is a very promising energyconversion process. However, it requires new advances in order to

approach the record efficiencies achieved by state-o-the-art semi-conductor solar cells. Several aspects directly related to the natureo these devices need to be considered, including the materialsused to abricate the plasmonic nanostructures, the semiconduc-tors used to capture the photoexcited hot electrons, their archi-tectures and even the abrication method. Tis section reviewsand discusses some o the most important aspects that need majoradvances to urther improve the efficiency o this energy conver-sion method.

Efficiency and fabrication cost of plasmonic versus conven-tional photovoltaic devices. A very important point that needs tobe careully considered is the efficiency o the plasmonic devicesrelative to that o conventional photovoltaic cells. Currently, themost commonly used solar cells are based on crystalline Si (c-Si).

SB

EF

EF

Ev

Ec

Ev

Ec

Eg

=3.2eV

Eg

=3.2eV

Eg

=2.7

6eV

Eg

=2.5

eV

EF

Au

Ag Cu

AZO

ITO FTO GZO

TiO2 ZnO CeO2

AgBr

Metals Conducting oxidesWide-bandgapsemiconductors

Energy(Vvs.vac

uum)

Energy(Vvs.N

HE)

b

a

Solarspectralirradian

ceAM1.5(Wm

2n

m1)

500 1,500 2,5001,000 2,000

Wavelength (nm)

4

5

6

7

1

0

1

2

3

e

h+

Ext

inction

AZOAl

ITO

Cu

RuO2

Ag

Au

n = 4.0 1020cm3

n = 8.0 1020cm3

n = 10 1020cm3

n = 2.0 1020cm3

0.0

1.0

0.0

0.5

1.0

1.5

Figure 4 |Materials for plasmonic solar cells. a, Optical extinction of

metal (Al, Ag, Au and Cu) and conducting oxides (RuO 2, AZO and ITO)

plasmonic nanostructures. The solar irradiance spectrum is plotted in

the background. b, Fermi level (EF), conduction (Ec) and valence (Ev)

bands energy for selected metals, conducting oxides and wide-bandgap

semiconductors of interest for the fabrication of plasmonic nanostructures.

The energies are plotted on a potential scale (V) versus the normal

hydrogen electrode (NHE) and also versus vacuum. The values for Ecand

Evfor TiO2, ZnO, CeO2and AgBr were obtained from refs 19, 9, 100 and 101,

respectively. The EFvalues for Au, Ag, Cu, ITO, FTO, AZO and GZO were

obtained from refs 64, 4, 93 and 102105, respectively.

REVIEW ARTICLENATURE PHOTONICS DOI: 10.1038/NPHOTON.2013.238


6/9


Commercial solar cells generally have average efficiencies inthe range 1620%, although efficiencies has high as 28.3% haverecently been achieved in single-junction solar cells85. Second-generation photovoltaic devices consisting o thin-film cells maderom amorphous silicon, Cde or copperindiumgallium disele-nide (CIGS) deposited on cheap substrates such as plastic, glass orstainless steel have achieved efficiencies o up to 20.3% (re. 85).Tird-generation devices encompass a large range o new technol-ogies, including multijunction cells, multiband cells, hot-carrier

cells and solar thermal technologies86. Because they employ sev-eral semiconductors with different bandgaps, multijunction solarcells have considerably broader absorption spectra than single-junction solar cells. Tis approach recently resulted in a recordefficiency o 43.5% being realized under concentrated illumina-tion85. In contrast, the maximum efficiency achieved to date indye-sensitized solar cells based on hot-electron generation is only11.4% (re. 85). In contrast, plasmonic solar cells based on hot-electron generation are still in their inancy; their efficiencies arediscussed below.

Te energy rom surace plasmon decay is transerred to elec-trons over the whole DOS distribution o the nanostructuresconduction band. Tis creates a very broad distribution o hot-electron states above the Fermi energy, so that many hot electrons

may have insufficient energy to be injected into the semicon-ductor. White and Catchpole25 estimated the efficiency limit oSchottky-barrier-based plasmonic energy conversion devices.Assuming a parabolic DOS or the conduction band o metalssuch as Ag and Au, a Schottky barrier energy o SB= 1.2 eV, andan equal photoexcitation probability or all the electrons in theconduction band, results in a maximum IPCE o 8%. However,photoemission experiments in noble metals indicate that elec-trons close to the Fermi energy are preerentially excited overlower energy electrons26, which considerably increases the effi-ciency o this process. o reflect this observation, White andCatchpole25 also perormed calculations or a more realisticmodel in which the effective conduction band edge is 0.15 eVbelow the Fermi level. Such a narrow DOS distribution makes

the emission efficiency approach a step-like unction in which theprobability o hot electrons with energies higher than SBcross-ing the barrier approaches 100%. For this case, they calculated aconsiderably higher efficiency o 22.6% or SB= 1.4 eV. Furtherengineering o the DOS o the plasmonic material to avour pre-erential excitation o electronic levels close to the Fermi energy,such as in semiconductor-based plasmonic nanostructures, alongwith tuning o the Schottky barrier, will a llow efficiencies well inexcess o 22% (re. 25). Furthermore, the use o materials withLSPR requencies spread over the whole solar spectrum will ena-ble more efficient utilization o the solar radiation, as shown inthe next subsection.

Efficiencies higher than 22% certainly make plasmonic solarcells an interesting alternative. However, the success o this tech-

nology is dependent on realizing a high efficiency while keepingabrication costs low. Te cost o Si has dropped over the past ewyears, but c-Si solar cells require relatively thick (180300 m)films with their associated expensive processing 87. Second- andthird-generation thin-film-based solar cells are cheaper to ab-ricate, but their materials have poor light trapping, so that plas-monic and dielectric light-trapping mechanisms are required3. Inthis sense, plasmonic solar cells combine the low abrication costso second-generation solar cells with the highly efficient light trap-ping o their plasmonic elements.

Te method used or abricating plasmonic energy conversiondevices also greatly affects their cost. So ar, they have been mostlyabricated by inexpensive solution-based methods in which com-mercially available noble metal and semiconductor nanoparti-cles are properly combined5,34,65,72. Noble metal nanoparticles o

controlled size have also been obtained by photocatalysis6,37,63,70,electrodeposition35,74, electrodeposition through a thin aluminananomask73, electrostatic sel-assembly75 and by a modifiedurkevich method33,88. More complicated structures, such as coreshell nanoparticles, have been abricated60,61. Also, noble metalnanoparticle-decorated iO2 nanotubes have been abricated bythe hydrothermal method45 and electrochemical anodic oxida-tion48,49. Only a ew studies have used lithography to achieve moreelaborate plasmonic nanostructure shapes, such as nanorods10,29

and nanowires12. Although the above-mentioned techniques areuseul or demonstrating this energy conversion method, scalabletechniques that provide a high degree o homogeneity over largerareas together with fine control o the sizes and shapes o nano-structures will be needed to urther develop and eventually com-mercialize these kinds o devices. o realize these goals, techniquesbased on physical vapour deposition that have been modified toenable nanoparticles to be created rom a high-pressure plasmaare promising8991. Tey allow reactive gases such as oxygen tobe added, enabling the ormation o size-controlled nanoparti-cles o oxides such as iO2, aluminium-doped zinc oxide (AZO)and IO. Other emerging technologies, such as nanoimprinting,should also be considered or this application92. Te use o suchtechniques will lead to important breakthroughs in the develop-

ment o this technology.

Plasmonic materials and semiconductor acceptors for opti-mized solar absorption. So ar, noble metals such as Ag and Auhave been virtually the only materials used in plasmonic energyconversion. Noble-metal nanostructures exhibit very intenseLSPR excitations, which give rise to very strong light absorption,mainly in the visible region. Te requency at which plasmonicnanostructures undergo LSPR strongly depends on their carrierconcentration. For spherical nanoparticles in air, LSPR occurs at arequency o LSPRp/3, where the plasma requency pdirectlydepends on the carrier concentration neaccording to67

(1)m*0

nee2

p=

where e is the charge o an electron, m* is the effective mass oan electron and 0 is the permittivity o ree space. Figure 4ashows the calculated optical extinction (that is, optical absorp-tion) or Al, Ag, Au and Cu nanoparticles superimposed on thesolar spectral irradiance AM1.5. Tese nanoparticles have veryintense but narrow absorptions in the lower wavelength regiono the visible range because o their relatively high carrier con-centrations93(around 5.9 1022cm3or Ag and Au). However, asignificant portion o the solar radiation that reaches the Earthssurace has longer wavelengths; about 40% o the solar radia-tion has a wavelength longer than 800 nm (re. 29). It is there-

ore o paramount importance to find materials that allow us touse a greater portion o the solar spectrum. Conducting oxidesand semiconductors have lower carrier concentrations thannoble metals, and hence lower plasma requencies, placing theirresonances in the near-inrared range. Figure 4a shows the cal-culated extinctions o the widely used conducting oxides RuO2,aluminium-doped zinc oxide (AZO) and IO. Conducting oxideshave broader absorptions than noble metals, mainly because otheir higher optical losses94. Te use o plasmonic nanostruc-tures made rom conducting oxides to capture and convert theinrared region o the solar spectrum will extend the range oapplication o plasmonic solar cells, significantly increasing theiroverall efficiency.

Furthermore, it is possible to use the doping concentrationin conducting oxides such as AZO and IO to tune their carrier



7/9


concentration, and hence the spectral region in which plasmonicresonance takes place. For example, the carrier concentration oAZO can be varied in the range 0.510 10 20cm3by changing theconcentration o Al; this enables the surace plasmon resonance tobe varied over the wide range o 2,200880 nm (re. 95), as shownin Fig. 4a. In addition, it is possible to create alloys o two or moremetals with fine-tuned plasmonic properties96. Nanostructuresmade rom noble-metaltransition-metal alloys are the best can-didates or this application because their Fermi level and surace

plasma requency LSPR can be modified. Other materials, suchas conducting transition-metal nitrides (or example, iN, ZrN,HN and aN) are interesting or plasmonic applications as theyexhibit metallic properties at visible requencies97,98. Analogous tomultiple-junction solar cells in which semiconductors with di-erent bandgaps are combined in order to cover the whole solarspectrum, systems that combine multiple metals and conductingoxides will permit the absorption spectrum o the devices to bematched to the solar spectrum.

Te semiconductor used to trap the photoexcited hot electronsgreatly affects the charge injection mechanism. Important actorsto consider are the bandgap o the semiconductor, which affectsthe height o the semiconductormetal Schottky barrier, and thedensity o available states in the conduction band, which affects

the efficiency o the hot-electron injection process. Good align-ment o the Fermi level o the plasmonic nanostructures with thebands o the semiconductor is important to avour efficient carrierinjection99. As shown in Fig. 4b, the Fermi energy level is around0 V on the normal hydrogen electrode scale9or noble metals andconducting oxides. Te energy at which surace plasmons transerto hot electrons on decay has been shown to depend strongly onthe material, and the size and shape o the plasmonic nanostruc-tures; it typically ranges between 1 eV and 4 eV or Au and Agnanostructures9,20,27. Tus, the semiconductor needs to be cho-sen such that the photoexcited hot electrons can overcome theSchottky barrier SB. Figure 4b depicts the positions o the conduc-tion and valence bands o several wide-bandgap semiconductorso interest or plasmonic energy conversion. iO2is by ar the most

requently used semiconductor material. It has a wide bandgap(Eg = 3.3 eV) and an excellent electron-accepting ability becauseo the high DOS in its conduction band. Tis is mainly becauseo the d-orbital nature o its conduction band; this contrasts withother typical metal oxides (such as ZnO, SnO 2and In2O3) whoseconduction bands are mainly composed o the s or sp orbital othe metal atoms11. Tese characteristics make iO2 an excellentcandidate or this application. Nevertheless, other semiconductorssuch as ZnO, CeO2and AgBr have suitably positioned conductionand valence bands, and they can also serve as efficient electronacceptors. Further comparative studies are needed to determinethe most suitable material based on the particular characteristicso the plasmonic nanostructures used.

Conclusions and outlookGeneration o hot electrons in plasmonic nanostructures is a verypromising energy conversion mechanism, and it has ound inter-esting applications in photovoltaic and photocatalytic devices.Tis review has covered the undamentals o hot-electron gen-eration, injection and regeneration in plasmonic nanostructures.Although still in its inancy, this field has witnessed considerableprogress, offering potentially high energy conversion efficiencieswhile keeping abrication costs low. Realizing ast hot-electroninjection beore recombination and optimal carrier regenerationare key actors that will allow higher conversion efficiencies to beobtained. In addition, the use o new plasmonic materials suchas semiconductors and conducting oxides is proposed to extendthe spectral range o light absorption and hot-electron generation,allowing optimum utilization o the solar spectrum.

Received 9 May 2013; accepted 2 August 2013; published30 January 2014

References1. Atwater, H. A. Te promise o plasmonics. Sci. Am.296,5662 (2007).2. Mendes, M. J., Luque, A., obas, I. & Mart, A. Plasmonic light enhancement

in the near-field o metallic nanospheroids or application in intermediate bandsolar cells.Appl. Phys. Lett.95,071105 (2009).

3. Atwater, H. A. & Polman, A. Plasmonics or improved photovoltaic devices.Nature Mater.9,205213 (2010).

4. Zhao, G., Kozuka, H. & Yoko, . Solgel preparation and photoelectrochemicalproperties o iO2films containing Au and Ag metal particles. Tin Solid Films277,147154 (1996).

5. ian, Y. & atsuma, . Plasmon-induced photoelectrochemistry at metalnanoparticles supported on nanoporous iO2. Chem. Commun.18101811 (2004).

6. ian, Y. & atsuma, . Mechanisms and applications o plasmon-inducedcharge separation at iO2films loaded with gold nanoparticles.J. Am. Chem.Soc.127,76327637 (2005).

7. Knight, M. W., Sobhani, H., Nordlander, P. & Halas, N. J. Photodetection withactive optical antennas. Science332,702704 (2011).

8. Wang, F. & Melosh, N. A. Plasmonic energy collection through hot carrierextraction. Nano Lett.11,54265430 (2011).

9. Linic, S., Christopher, P. & Ingram, D. B. Plasmonic-metal nanostructures orefficient conversion o solar to chemical energy. Nature Mater.10,911921 (2011).

10. Nishijima, Y.et al.Near-inrared plasmon-assisted water oxidation.J. Phys.

Chem. Lett.3,12481252 (2012).11. Du, L., Furube, A., Hara, K., Katoh, R. & achiya, M. Ultraast plasmon

induced electron injection mechanism in goldiO2nanoparticle system.J. Photochem. Photobiol. C15,2130 (2013).

12. Knight, M. W.et al.Embedding plasmonic nanostructure diodes enhances hotelectron emission. Nano Lett.13,16871692 (2013).

13. Mubeen, S.et al.An autonomous photosynthetic device in which all chargecarriers derive rom surace plasmons. Nature Nanotechnol.8,247251 (2013).

14. Semenov, A. D., Goltsman, G. N. & Sobolewski, R. Hot-electron effect insuperconductors and its applications or radiation sensors. Supercond. Sci.echnol.15,R1 (2002).

15. Hertz, H. Ueber einen Einfluss des ultravioletten Lichtes au die electrischeEntladung.Ann. Phys.267,9831000 (1887).

16. Gadzuk, J. W. On the detection o chemically-induced hot electrons in suraceprocesses: rom X-ray edges to Schottky barriers.J. Phys. Chem. B106,82658270 (2002).

17. Nienhaus, H. Electronic excitations by chemical reactions on metal suraces.Surf. Sci. Rep.45,178 (2002).18. ORegan, B. & Grtzel, M. A low-cost, high-efficiency solar cell based on dye-

sensitized colloidal iO2films. Nature353,737740 (1991).19. Grtzel, M. Photoelectrochemical cells. Nature414,338344 (2001).20. Snnichsen, C.et al.Drastic reduction o plasmon damping in gold nanorods.

Phys. Rev. Lett.88,077402 (2002).21. Homann, J. & Steinmann, W. Plasma resonance in the photoemission o silver.

Phys. Status Solidi B30, K53K56 (1968).22. Endriz, J. G. & Spicer, W. E. Surace-plasmon-one-electron decay and its

observation in photoemission. Phys. Rev. Lett.24,6468 (1970).23. Lehmann, J.et al.Surace plasmon dynamics in silver nanoparticles studied by

emtosecond time-resolved photoemission. Phys. Rev. Lett.85,29212924 (2000).

24. Inagaki, ., Kagami, K. & Arakawa, E. . Photoacoustic observation ononradiative decay o surace plasmons in silver. Phys. Rev. B24,36443646 (1981).

25. White, . P. & Catchpole, K. R. Plasmon-enhanced internal photoemission orphotovoltaics: theoretical efficiency limits.Appl. Phys. Lett.101,073905 (2012).

26. Berglund, C. N. & Spicer, W. E. Photoemission studies o copper and silver:experiment. Phys. Rev.136,A1044A1064 (1964).

27. Rycenga, M.et al.Controlling the synthesis and assembly o silvernanostructures or plasmonic applications. Chem. Rev.111,36693712 (2011).

28. Moskovits, M. Hot electrons cross boundaries. Science332,676677 (2011).29. Nishijima, Y., Ueno, K., Yokota, Y., Murakoshi, K. & Misawa, H. Plasmon-

assisted photocurrent generation rom visible to near-inrared wavelengthusing a Au-nanorods/iO2electrode.J. Phys. Chem. Lett.1,20312036 (2010).

30. Ohko, Y.et al.Multicolour photochromism o iO2films loaded with silvernanoparticles. Nature Mater.2,2931 (2003).

31. Naoi, K., Ohko, Y. & atsuma, . iO2films loaded with silver nanoparticles:control o multicolor photochromic behavior.J. Am. Chem. Soc.126,36643668 (2004).

32. Hirakawa, . & Kamat, P. V. Photoinduced electron storage and surace

plasmon modulation in Ag@iO2clusters. Langmuir20,56455647 (2004).



8/9


33. Lana-Villarreal, . & Gmez, R. uning the photoelectrochemistry onanoporous anatase electrodes by modification with gold nanoparticles:development o cathodic photocurrents. Chem. Phys. Lett.414,489494 (2005).

34. Yu, K., ian, Y. & atsuma, . Size effects o gold nanaoparticles on plasmon-induced photocurrents o goldiO2nanocomposites. Phys. Chem. Chem. Phys.8,54175420 (2006).

35. Sakai, N., Fujiwara, Y., akahashi, Y. & atsuma, . Plasmon-resonance-basedgeneration o cathodic photocurrent at electrodeposited gold nanoparticlescoated with iO2films. ChemPhysChem10,766769 (2009).

36. Kowalska, E., Abe, R. & Ohtani, B. Visible light-induced photocatalytic reactiono gold-modified titanium(IV) oxide particles: action spectrum analysis. Chem.Commun.241243 (2009).

37. oyoda, ., sugawa, S. & Shen, Q. Photoacoustic spectra o Au quantumdots adsorbed on nanostructured iO2electrodes together with thephotoelectrochemical current characteristics.J. Appl. Phys.105,034314 (2009).

38. Gomes Silva, C., Jurez, R., Marino, ., Molinari, R. & Garca, H. Influence oexcitation wavelength (UV or visible light) on the photocatalytic activity otitania containing gold nanoparticles or the generation o hydrogen or oxygenrom water.J. Am. Chem. Soc.133,595602 (2010).

39. Kowalska, E., Mahaney, O. O. P., Abe, R. & Ohtani, B. Visible-light-inducedphotocatalysis through surace plasmon excitation o gold on titania suraces.Phys. Chem. Chem. Phys.12,23442355 (2010).

40. Ide, Y., Matsuoka, M. & Ogawa, M. Efficient visible-light-inducedphotocatalytic activity on gold-nanoparticle-supported layered titanate.J. Am.Chem. Soc.132,1676216764 (2010).

41. Valverde-Aguilar, G., Garca-Macedo, J. A., Rentera-apia, V. & Aguilar-Franco, M. Photoconductivity studies on amorphous and crystalline iO2filmsdoped with gold nanoparticles.Appl. Phys. A103,659663 (2011).

42. Ingram, D. B., Christopher, P., Bauer, J. L. & Linic, S. Predictive model or thedesign o plasmonic metal/semiconductor composite photocatalysts.ACSCatal.1,14411447 (2011).

43. anaka, A.et al.Goldtitanium(IV) oxide plasmonic photocatalysts preparedby a colloid-photodeposition method: correlation between physical propertiesand photocatalytic activities. Langmuir28,1310513111 (2012).

44. Shi, X., Ueno, K., akabayashi, N. & Misawa, H. Plasmon-enhancedphotocurrent generation and water oxidation with a gold nanoisland-loaded titanium dioxide photoelectrode.J. Phys. Chem. C117,24942499 (2013).

45. Gong, D.et al.Silver decorated titanate/titania nanostructures or efficient solardriven photocatalysis.J. Solid State Chem.189,117122 (2012).

46. Sakai, N., Sasaki, ., Matsubara, K. & atsuma, . Layer-by-layer assembly ogold nanoparticles with titania nanosheets: control o plasmon resonance and

photovoltaic properties.J. Mater. Chem.20,43714378 (2010).47. Shiraishi, Y.et al.Platinum nanoparticles supported on anatase titanium

dioxide as highly active catalysts or aerobic oxidation under visible lightirradiation.ACS Catal.2,19841992 (2012).

48. Wu, F.et al.Photocatalytic activity o Ag/iO2nanotube arrays enhanced bysurace plasmon resonance and application in hydrogen evolution by watersplitting. Plasmonics8,501508 (2012).

49. Chen, K.et al.Effect o Ag nanoparticle size on the photoelectrochemicalproperties o Ag decorated iO2nanotube arrays.J. Alloys Compd554,7279 (2013).

50. Chen, Z. H.et al.Vertically aligned ZnO nanorod arrays sentisized with goldnanoparticles or Schottky barrier photovoltaic cells.J. Phys. Chem. C113,1343313437 (2009).

51. Chen, H. M.et al.Plasmon inducing effects or enhanced photoelectrochemicalwater splitting: X-ray absorption approach to electronic structures.ACS Nano6,73627372 (2012).

52. Han, Z.et al.Visible-light photocatalytic application o hierarchical Au-ZnOflower-rod heterostructures via surace plasmon resonance. Plasmonics8,11931202 (2013).

53. Primo, A., Marino, ., Corma, A., Molinari, R. & Garca, H. Efficient visible-light photocatalytic water splitting by minute amounts o gold supportedon nanoparticulate CeO2obtained by a biopolymer templating method.J. Am. Chem. Soc.133,69306933 (2011).

54. Kominami, H., anaka, A. & Hashimoto, K. Mineralization o organic acidsin aqueous suspensions o gold nanoparticles supported on cerium(IV) oxidepowder under visible light irradiation. Chem. Commun.46,12871289 (2010).

55. Chen, X., Li, P., ong, H., Kako, . & Ye, J. Nanoarchitectonics o a Aunanoprism array on WO3film or synergistic optoelectronic response. Sci.echnol. Adv. Mater.12,044604 (2011).

56. Zhou, X., Hu, C., Hu, X., Peng, . & Qu, J. Plasmon-assisted degradation otoxic pollutants with AgAgBr/Al2O3under visible-light irradiation.J. Phys.Chem. C114,27462750 (2010).

57. Hu, C.et al.Plasmon-induced photodegradation o toxic pollutants with AgAgI/Al2O3under visible-light irradiation.J. Am. Chem. Soc.132,857862 (2009).

58. Wang, P.et al.Ag@AgCl: a highly efficient and stable photocatalyst activeunder visible light.Angew. Chem.47,79317933 (2008).

59. Kochuveedu, S. ., Kim, D.-P. & Kim, D. H. Surace-plasmon-inducedvisible light photocatalytic activity o iO2nanospheres decorated byAu nanoparticles with controlled configuration.J. Phys. Chem. C116,25002506 (2012).

60. Zhang, N., Liu, S., Fu, X. & Xu, Y.-J. Synthesis o M@iO 2(M = Au, Pd, Pt)core-shell nanocomposites with tunable photoreactivity.J. Phys. Chem. C115,91369145 (2011).

61. Zheng, Z.et al.Facile in situsynthesis o visible-light plasmonic photocatalystsM@iO2(M = Au, Pt, Ag) and evaluation o their photocatalytic oxidation obenzene to phenol.J. Mater. Chem.21,90799087 (2011).

62. Furube, A., Du, L., Hara, K., Katoh, R. & achiya, M. Ultraast plasmon-induced electron transer rom gold nanodots into iO2nanoparticles.J. Am.Chem. Soc.129,1485214853 (2007).

63. ian, Y., Wang, X., Zhang, D., Shi, X. & Wang, S. Effects o electron donorson the perormance o plasmon-induced photovoltaic cell.J. Photochem.Photobiol., A199,224229 (2008).

64. ian, Y., Shi, X., Lu, C., Wang, X. & Wang, S. Charge separation in solid-stategold nanoparticles-sensitized photovoltaic cell. Electrochem. Commun.11,16031605 (2009).

65. Du, L.et al.Plasmon-induced charge separation and recombination dynamicsin goldiO2nanoparticle systems: dependence on iO2particle size.J. Phys.Chem. C113,64546462 (2009).

66. Inouye, H., anaka, K., anahashi, I. & Hirao, K. Ultraast dynamics ononequilibrium electrons in a gold nanoparticle system. Phys. Rev. B57,1133411340 (1998).

67. Kreibig, U. & Vollmer, M. Optical Properties of Metal Clusters. (Springer, 1995).68. Langhammer, C., Yuan, Z., Zori, I. & Kasemo, B. Plasmonic properties o

supported Pt and Pd nanostructures. Nano Lett.6,833838 (2006).69. Skoplaki, E. & Palyvos, J. A. On the temperature dependence o photovoltaic

module electrical perormance: a review o efficiency/power correlations. Sol.Energy83,614624 (2009).

70. Kazuma, E., Sakai, N. & atsuma, . Nanoimaging o localized plasmon-induced charge separation. Chem. Commun.47,57775779 (2011).

71. Kazuma, E. & atsuma, . Photoelectrochemical analysis o allowed andorbidden multipole plasmon modes o polydisperse Ag nanorods.J. Phys.Chem. C 117,24353441 (2013).

72. Yu, K., Sakai, N. & atsuma, . Plasmon resonance-based solid-statephotovoltaic devices. Electrochem. Commun.76,161 (2008).

73. akahashi, Y. & atsuma, . Electrodeposition o thermally stable gold and

silver nanoparticle ensembles through a thin alumina nanomask. Nanoscale2,14941499 (2010).

74. akahashi, Y. & atsuma, . Solid state photovoltaic cells based on localizedsurace plasmon-induced charge separation.Appl. Phys. Lett.99,182110182113 (2011).

75. Reineck, P.et al.A solid-state plasmonic solar cell via metal nanoparticle sel-assembly.Adv. Mater.24,47504755 (2012).

76. Nakanishi, H.et al.Photoconductance and inverse photoconductance in filmso unctionalized metal nanoparticles. Nature460,371375 (2009).

77. Warren, S. C., Walker, D. A. & Grzybowski, B. A. Plasmoelectronics: couplingplasmonic excitation with electron flow. Langmuir28,90939102 (2012).

78. Mubeen, S., Hernandez-Sosa, G., Moses, D., Lee, J. & Moskovits, M. Plasmonicphotosensitization o a wide band gap semiconductor: converting plasmons tocharge carriers. Nano Lett.11,55485552 (2011).

79. Son, M.-S.et al.Surace plasmon enhanced photoconductance andsingle electron effects in mesoporous titania nanofibers loaded with gold

nanoparticles.Appl. Phys. Lett.96,023115 (2010).80. Lee, Y. K.et al.Surace plasmon-driven hot electron flow probed with metal-semiconductor nanodiodes. Nano Lett.11,42514255 (2011).

81. Lee, Y. K., Park, J. & Park, J. Y. Te effect o dye molecules and surace plasmonsin photon-induced hot electron flows detected on Au/iO2nanodiodes.J. Phys.Chem. C116,1859118596 (2012).

82. Wadell, C., Antosiewicz, . J. & Langhammer, C. Optical absorptionengineering in stacked plasmonic Au-SiO2-Pd nanoantennas. Nano Lett.12,47844790 (2012).

83. Garcia, G.et al.Dynamically modulating the surace plasmon resonance odoped semiconductor nanocrystals. Nano Lett.11,44154420 (2011).

84. Pradhan, A. K., Holloway, ., Mundle, R., Dondapati, H. & Bahoura, M. Energyharvesting in semiconductor-insulator-semiconductor junctions throughexcitation o surace plasmon polaritons.Appl. Phys. Lett.100,061127 (2012).

85. Green, M. A., Emery, K., Hishikawa, Y., Warta, W. & Dunlop, E. D. Solarcell efficiency tables (version 39). Prog. Photovoltaics Res. Appl.20,1220 (2012).



9/9


86. Green, M. A. Tird generation photovoltaics: solar cells or 2020 and beyond.Physica E14,6570 (2002).

87. Mokkapati, S. & Catchpole, K. R. Nanophotonic light trapping in solar cells.J. Appl. Phys.112,101101 (2012).

88. Atar, F. B.et al.Plasmonically enhanced hot electron based photovoltaic device.Opt. Express21,71967201 (2013).

89. Binns, C. Nanoclusters deposited on suraces. Surf. Sci. Rep.44,149 (2001).90. Palmer, R. E., Pratontep, S. & Boyen, H.-G. Nanostructured suraces rom size-

selected clusters. Nature Mater.2,443448 (2003).91. Wegner, K., Piseri, P., Vahedi-areshi, H. & Milani, P. Cluster beam deposition:

a tool or nanoscale science and technology.J. Phys. D 39,R439 (2006).92. Perelaer, J. & Schubert, U. S. Novel approaches or low temperature sintering

o inkjet-printed inorganic nanoparticles or roll-to-roll (R2R) applications.J. Mater. Res.28,564573 (2013).

93. Kittel, C. Introduction to Solid State Physics. (Wiley, 1976).94. Hvel, H., Fritz, S., Hilger, A., Kreibig, U. & Vollmer, M. Width o cluster

plasmon resonances: bulk dielectric unctions and chemical interace damping.Phys. Rev. B48,1817818188 (1993).

95. Mendelsberg, R. J.et al.Achieving high mobility ZnO:Al at very high growthrates by dc filtered cathodic arc deposition.J. Phys. D 44,232003 (2011).

96. West, P. R.et al.Searching or better plasmonic materials. Las. Photon. Rev.4,795808 (2010).

97. Naik, G. V., Kim, J. & Boltasseva, A. Oxides and nitrides as alternativeplasmonic materials in the optical range. Opt. Mater. Express1,10901099 (2011).

98. Naik, G. V.et al.itanium nitride as a plasmonic material or visible and near-inrared wavelengths. Opt. Mater. Express2,478489 (2012).

99. Cushing, S. K.et al.Photocatalytic activity enhanced by plasmonic resonantenergy transer rom metal to semiconductor.J. Am. Chem. Soc.134,1503315041 (2012).

100. Magesh, G., Viswanathan, B., Viswanath, R. P. & Varadarajan, . K.Photocatalytic behavior o CeO2-iO2system or the degradation o methyleneblue. Indian J. Chem. A48,480488 (2009).

101. af, E. A., Philipp, H. R. & Apker, L. Photoelectric emission rom the valenceband in AgBr. Phys. Rev. 110,876878 (1958).

102. Park, Y., Choong, V., Gao, Y., Hsieh, B. R. & ang, C. W. Work unctiono indium tin oxide transparent conductor measured by photoelectronspectroscopy.Appl. Phys. Lett.68,26992701 (1996).

103. Helander, M. G., Greiner, M. ., Wang, Z. B., ang, W. M. & Lu, Z. H. Workunction o fluorine doped tin oxide. J. Vac. Sci. echnol. A29,011019 (2011).

104. Wang, W. et al.Dependence o aluminum-doped zinc oxide work unctionon surace cleaning method as studied by ultraviolet and X-ray photoelectronspectroscopies.Appl. Surf. Sci.257,38843887 (2011).

105. Yoshida, K., Aoki, ., Suzuki, A., Matsushita, . & Okuda, M. Work unctioncontrolled transparent conductive films with zinc oxide system by pulsed laserdeposition method.J. Vac. Soc. Japan51,172174 (2008).

AcknowledgementsTe author thanks A. Anders and R. Mendelsberg or insightul discussions. Tis work

was supported by the Assistant Secretary or Energy Efficiency and Renewable Energy,

Office o Building echnology, o the US Department o Energy under Contract No.

DE-AC02-05CH11231.

Additional informationReprints and permissions inormation is available at www.nature.com/reprints.

Correspondence and requests or materials should be addressed to the author.

Competing financial interestsTe author declares no competing financial interests.


Documents

Clavero 2014 Nature Photonics