nmat3029

ARTICLESPUBLISHED ONLINE: 15 MAY 2011 | DOI: 10.1038/NMAT3029

Nanoantenna-enhanced gas sensing in a singletailored nanofocusNa Liu1†, Ming L. Tang1†, Mario Hentschel2†, Harald Giessen2 and A. Paul Alivisatos1*

Metallic nanostructures possess plasmonic resonances that spatially confine light on the nanometre scale. In the ultimate limitof a single nanostructure, the electromagnetic field can be strongly concentrated in a volume of only a few hundred nm3 or less.This optical nanofocus is ideal for plasmonic sensing. Any object that is brought into this single spot will influence the opticalnanostructure resonance with its dielectric properties. Here, we demonstrate antenna-enhanced hydrogen sensing at thesingle-particle level. We place a single palladium nanoparticle near the tip region of a gold nanoantenna and detect the changingoptical properties of the system on hydrogen exposure by dark-field microscopy. Our method avoids any inhomogeneousbroadening and statistical effects that would occur in sensors based on nanoparticle ensembles. Our concept paves the roadtowards the observation of single catalytic processes in nanoreactors and biosensing on the single-molecule level.

Nanoscale field confinement bymetallic nanostructures is oneof the key aspects of nanophotonics1–7. Localized surfaceplasmon resonances (LSPRs) of metallic nanostructures8–12,

which are associatedwith collective oscillations of free electrons, cangenerate large field confinement in an extremely small volume13.This so-called nanofocusing offers remarkable opportunities toimprove fluorescence14–16, Raman scattering17,18, single-moleculedetection19–21 and nonlinear nanospectroscopy22. Metallic nano-structures with sharp corners or edges are especially favourablefor these purposes. The local field strength in the vicinity of asharp gold tip can be enhanced by several orders of magnitudecompared with the incident light2,23. Importantly, nanofocusinghas immediate implications for plasmon-mediated chemistry24and sensing25–29. The strong enhancement of the local fields ina small sensing volume can be used to monitor the response ofLSPRs to a change in the surrounding dielectric permittivity8.So far, plasmonic sensing has been widely applied in manydifferent research areas including label-free detection of molecularbinding30 as well as monitoring low concentrations of biochemicalagents31,32 and gases33,34.

The detection of flammable gases such as hydrogen is a vitalsafety issue concomitant with the fast development of fuel celltechnology. Concentrations of hydrogen exceeding 4% may igniteexplosively. Palladium, one of themost catalytically active transitionmetals, can absorb a substantial quantity of hydrogen withinits crystal lattice and form palladium hydride in a reversiblemanner35. The electrical and dielectric properties of palladiumchange on hydrogen exposure36, which in turn form the backboneof electrical37,38 and optical hydrogen sensing39. When comparedwith sensors requiring electrical measurements, optical detection ofhydrogen is much more desirable owing to its inherent safety, asno sparking can occur.

Recently, palladium-based nanomaterials40–43 have attractedmuch attention in that they show faster reaction kinetics owingto their shorter diffusion lengths44. Nevertheless, the LSPRs ofpalladium nanoparticles possess very broad spectral resonanceprofiles45,46 resulting from interband transitions in the whole

1Department of Chemistry, University of California, Berkeley, and Materials Sciences Division, Lawrence Berkeley National Laboratory, Berkeley, California94720, USA, 24. Physikalisches Institut and Research Center SCoPE, Universität Stuttgart, D-70569 Stuttgart, Germany. †These authors contributedequally to this work. *e-mail: [email protected].

visible region47. This constitutes a fundamental obstacle for opticalhydrogen detection using palladium nanoparticles. Very recently,indirect optical sensing for hydrogen was introduced48,49, using anensemble of palladium nanoparticles which were separated fromunderlying gold discs by a SiO2 dielectric spacer. However, owingto the wide size distribution of the palladium particles and golddiscs, only inhomogeneously broadened average characteristicswere measured. Also, the presence of the SiO2 spacer leads to thedielectric screening of the near-fields of the gold discs.

In contrast to studies on an ensemble, single-entity measure-ments offer a unique tool for probing individual behaviour andprocesses, which cannot be distinguished on the ensemble level.Importantly, this advantage is favourable for plasmonic gas sensingand opens an avenue towards understanding the influence ofgeometries and nano-sizing on hydrogen storage performance50.Here we demonstrate resonant antenna-enhanced single-particlehydrogen sensing in the visible region. We present a fabrica-tion approach to precisely position a single palladium nanopar-ticle in the nanofocus of a gold nanoantenna (see Fig. 1b). Thestrongly enhanced gold-particle plasmon near-fields can sensethe change in the dielectric function of the proximal palladiumnanoparticle as it absorbs or releases hydrogen36. Light scatteredby the system is collected by a dark-field microscope with at-tached spectrometer and the LSPR change is read out in realtime. We demonstrate how the antenna enhancement effect canbe quantitatively controlled by manipulating the distance betweensensing object and nanoantenna. Additionally, we lay out howthe nanoantenna shape plays a key role for optimum sensitivity.Specifically, we show an unambiguous correlation between thesensor response and the antenna geometry. In contrast, the singlepalladium nanoparticle alone is barely visible in the microscope,and no useful spectral information can be derived (see Fig. 1a).Our work provides a general blueprint for amplifying sensingsignals at the single-particle level, eliminating the statistical andaverage characteristics inherent to ensemble measurements48,49. Inparticular, the inclusion of catalytic species such as palladium canalso offer advantages to in situ monitoring of catalytic events51,

NATURE MATERIALS | VOL 10 | AUGUST 2011 | www.nature.com/naturematerials 631

© 2011 Macmillan Publishers Limited. All rights reserved

http://www.nature.com/doifinder/10.1038/nmat3029

mailto:[email protected]

http://www.nature.com/naturematerials

ARTICLES NATURE MATERIALS DOI: 10.1038/NMAT3029

Pd

Wavelength

a

b

Scat

tere

d in

tens

ity

Δ

Wavelength

Scat

tere

d in

tens

ity

Pd

Au

λ

Figure 1 | Schematic representation of antenna-enhanced single-particlehydrogen sensing. a, Hydrogen sensing with a single palladiumnanoparticle. Hydrogen molecules and adsorbed hydrogen atoms areshown in red. The palladium nanoparticle scatters weakly, showing anextremely damped and broad spectrum. The palladium particle alonecauses a barely detectable change on hydrogen exposure. b, Hydrogensensing using a resonant antenna-enhanced scheme. The same palladiumnanoparticle is placed at the nanofocus of a gold antenna, which scattersmuch more strongly. Hydrogen absorption on the palladium particlechanges its complex dielectric function, which causes a resonance shift(1λ) of the gold antenna that can be optically detected.

as well as opportunities to study enhancements in the efficiencyof chemical conversion processes in electrochemistry and surfacescience52. Our work will advance the field of catalytic-activitystudies, which largely rely on fluorescence techniques53. Findingappropriate fluorophores is still a hurdle for many experimentsassociated with fluorescent detection of catalytic events. Our ap-proach, however, offers a promising alternative by employingthe change of dielectric permittivity, that is, refractive index andabsorption, of targeted biochemical entities for catalytic-activitydetection and biosensing.

Figure 2 shows the simulated electric-field distributions of twopalladium–gold antennas at resonance. The plasmonic antennasare a gold triangle and a gold rod, respectively (see Fig. 2). Theincident light is polarized parallel to the longer axis of the system.It is evident that, when the gold antenna is resonantly excited,collective dipolar-like particle plasmons lead to greatly enhancedfields at the tips of the gold triangle or at the corners of the goldrod. Specifically, owing to the sharp tip feature of the gold triangle,the near-fields are spatially concentrated in a single spot right nextto the palladium particle.

The nanostructures with two materials on the same planewere manufactured by high-resolution electron-beam lithographyusing a dual-exposure method. Figure 3 shows the processingscheme. The gold antenna and the palladium particle weremanufactured on a glass substrate in two separate nanofabricationcycles, which are associated with electron-beam exposure, metalevaporation and lift-off (see Methods). The relative positions ofthe two metallic constituents were accurately aligned using goldmarkers54. Our fabrication approach offers a unique opportunityto independently control the sizes, positions and materials ofdifferent constituents.

The sample was placed in a stainless-steel chamber with opticallytransparent windows at room temperature. The chamber wasconnected to two gas channels (ultrahigh-purity hydrogen andnitrogen). The concentration of hydrogen was adjusted by tuningthe proportion of the two gases through mass-flow controllers.

E

10.9

(V m

¬1 )

¬10.9

2.2

¬2.0

0.4

¬0.4

20.4

(V m

¬1 )

¬20.4

9.3

¬8.8

3.4

0

¬3.1

0

E

a

b

Figure 2 | Electromagnetic finite-difference time-domain simulation ofthe local electric fields. The polarization is along the vertical direction. a, Agold triangle antenna with a palladium nanoparticle. Near the tip of the goldtriangle antenna, the electromagnetic fields are highly localized andstrongly enhanced. Any change of the permittivity of the palladium particlewill significantly influence the plasmonic resonance of the gold antenna.The side length of the equilateral gold triangle is 110 nm and the goldthickness is 40 nm. The diameter of the palladium particle is 60 nm and thepalladium thickness is 40 nm. The gap distance between the gold antennaand the palladium particle is 10 nm. b, A gold rod antenna with a palladiumnanoparticle. The field enhancement in a rod antenna is more concentratedat the corners. The length and width of the gold rod are 130 nm and 80 nm,respectively, and its thickness is 40 nm. The diameter and thickness of thepalladium particle are 60 nm and 40 nm, respectively. The gap distancebetween the gold antenna and the palladium particle is 20 nm.

Figure 3b shows a dark-field image of the palladium–gold antennas.The distance between neighbouring structures is 20 µm, allowingfor isolated single-particle measurements. One representativescanning electron microscopy (SEM) image of the palladium–goldantennas is also shown in Fig. 3b, demonstrating the excellentalignment of the palladium nanoparticle with respect to thegold antenna. The side length of the equilateral gold triangleis 110 nm and the gold thickness is 40 nm. The diameter ofthe palladium particle is 60 nm and the palladium thickness is40 nm. The grain-like features on the substrate result from athin metal film evaporated after all the optical measurementswere done for SEM imaging of the nanostructures on thenonconductive glass substrate.

Figure 4 shows the scattering spectra of the gold triangle sensorin response to hydrogen exposure. To investigate the influence

632 NATURE MATERIALS | VOL 10 | AUGUST 2011 | www.nature.com/naturematerials




NATURE MATERIALS DOI: 10.1038/NMAT3029 ARTICLESFirst exposure

Au evaporationa

b

Dark-field image

20 µm 100 nm

SEM image

Second exposurePd evaporation

Lift-off Lift-off

Figure 3 | Manufacturing process. a, The fabrication of a planar nanostructure with hybrid materials using double electron-beam lithography incombination with a double lift-off procedure. The relative positions of the gold antenna and the palladium particle were accurately aligned using goldmarkers, which are represented by the three small gold dots in the schematic diagram. b, Left: Dark-field image of the palladium–gold antennas. Right:Representative SEM image of a single palladium–gold antenna.

Gap distance = 10 nm

0

33

16

0

16

8

0

Gap distance = 70 nm Gap distance = 90 nm

500 550 600Wavelength (nm)

Wavelength (nm)

650 700

636 640 644 648

500 550 600Wavelength (nm)

650 700 500 550 600Wavelength (nm)

650 700

Scat

tere

d in

tens

ity (

a.u.

)

Pres

sure

(to

rr)

Scat

tere

d in

tens

ity (

a.u.

)

Scat

tere

d in

tens

ity (

a.u.

)

Cyc

le 1

Cyc

le 1

Cyc

le 2

Cyc

le 2

0

33

16

0

16

8

0

Wavelength (nm)620 624 628 632


Pres

sure

(to

rr)

16

8

0

Pres

sure

(to

rr)

Cyc

le 1

Cyc

le 2

Cyc

le 1 Cyc

le 1

Cyc

le 2

0 torr0 torr

33 torr

33 torr

16 torr16 torr

16 torr 16 torr

8 torr 8 torr

0 torr0 torr

0 torr0 torr

16 torr

8 torr

0 torr

a b c

d = 10 nm d = 70 nm d = 90 nm

Figure 4 | Optical-scattering measurements of a single palladium–gold triangle antenna on hydrogen exposure in dependence on separation d betweenthe gold antenna and the palladium particle. a, d= 10 nm. b, d= 70 nm. c, d=90 nm. In each part, the hydrogen partial pressure is raised from 0 torr tohigher pressures and driven back to 0 torr. Two cycles are shown. 33 torr corresponds approximately to a 4% hydrogen concentration. The vertical bars ineach figure are guides to the eye. The right diagram in each part shows the behaviour of the resonance peak on hydrogen cycles one and two. The linesconnecting the data points are guides to the eye. The arrow indicates the direction of measurements. Our error bar for the peak determination is estimatedto be±1 nm, as shown in a, in the case of pronounced scattering spectra with a gold nanoantenna present. The SEM image of the nanostructure is shown ineach part accordingly. The scale bar is 50 nm. At larger distances, the spectral shift substantially decreases. The spectral shift does not go back to itsoriginal value owing to hysteresis effects of the hydrogen uptake.

of the local field enhancement on the sensor activity, the gapdistance d between the gold antenna and the palladium particleis systematically increased. Figure 4a shows the optical scatteringmeasurements of the gold triangle sensor with d as small as 10 nm.The spectral peak positions were extracted from the experimentalspectra. Smoothened curves are represented by black solid lines inthe same figure. To highlight the spectral shift in dependence on thehydrogen concentration, the extracted spectral peak positions areshown in the right columnof Fig. 4a.When the hydrogen pressure is0 torr, that is, pure nitrogen, a plasmonic resonance is visible around638 nm, which is represented by the blue curve in Fig. 4a. Thisresonance is associated with the resonant excitation of dipolar-likeparticle plasmons in the gold triangle (see Fig. 2a). When thehydrogen pressure is increased to 8 torr (∼1% concentration), a

clear resonance red-shift (∼5 nm) is observable and simultaneouslythe resonance intensity is dramatically suppressed (see the greencurve in Fig. 4a). As a result, the hydrogen absorption event inthe palladium particle is indirectly detected by monitoring thegold-antenna LSPR response. This relies on the fact that thedielectric permittivity of the palladium is modified by Fermi-levelshifting owing to incorporation of free electrons from the adsorbedhydrogen36,55. A subsequent pressure increase of hydrogen (16 torr,∼2%) leads to a further resonance red-shift (∼9 nm), as shown bythe red curve in Fig. 4a.

To probe the hydrogen desorption response, the hydrogenpressure is driven back to 0 torr in the second cycle by purgingwith nitrogen. As shown by the light-blue curve in Fig. 4a, theresonance peak does not return to the original position at 0 torr in






a b cGap distance = 20 nm

0

33

16

0

16

8

0

550 600Wavelength (nm)

Wavelength (nm)

d = 20 nm d = 50 nm d = 85 nm650 700

676 680 684 688Wavelength (nm)652 656 660 664


Scat

tere

d in

tens

ity (

a.u.

)

Pres

sure

(to

rr)

0

33

16

0

16

8

0

Pres

sure

(to

rr)

16

8

0

Pres

sure

(to

rr)

Cyc

le 1

Cyc

le 1

Cyc

le 2



650 700

Scat

tere

d in

tens

ity (

a.u.

)



650 700

Scat

tere

d in

tens

ity (

a.u.

)

Cyc

le 1

Cyc

le 1

Cyc

le 2

Cyc

le 2

Cyc

le 1

Cyc

le 2

0 torr

33 torr

16 torr

0 torr

33 torr

16 torr

16 torr

8 torr

0 torr

0 torr

16 torr

8 torr

0 torr

0 torr

16 torr

8 torr

0 torr

Figure 5 | Optical scattering measurements of a single palladium–gold rod antenna on hydrogen exposure in dependence on the separation d betweenthe gold antenna and the palladium particle. a, d= 20 nm. b, d= 50 nm. c, d=85 nm. In each part, the hydrogen partial pressure is raised from 0 torr tohigher pressures and driven back to 0 torr. Two cycles are shown. The vertical bars in each part are guides to the eye. The right diagram in each part showsthe behaviour of the resonance peak on hydrogen cycles one and two. The lines connecting the data points are guides to the eye. The arrow indicates thedirection of measurements. The SEM image of the nanostructure is shown in each part accordingly. The scale bar is 50 nm. At larger distances, the spectralshift substantially decreases. Hysteresis is also clearly observable.

the first cycle. It exhibits a net red-shift of around 4 nm, showinghysteresis. In fact, at low pressures the incorporation of hydrogentakes place at interstitial sites of the palladium lattice (α-phase).At high pressures46, a palladium hydride phase (β-phase) graduallybecomes stable. Hydrogen atoms progressively occupy lattice sitesand induce lattice expansion. The hysteresis is attributed to thelattice strain resulting from hydrogen incorporation41,49,50,56 and/orresidual hydrogen that cannot be removed from the Pd nanoparticleat room temperature57. This is the first optical observation ofhysteresis effects in a single palladium nanoparticle.

When the hydrogen pressure is increased to 16 torr in thesecond cycle, the resonance exhibits a red-shift (∼3 nm), yetmuch smaller than the 9 nm red-shift in the first cycle at thesame pressure. This reveals that hysteresis causes a deteriorationof the sensor behaviour. This could be improved or avoided byco-depositing nickel58 together with palladium for future practicalapplications. When the hydrogen pressure is increased to 33 torr(∼4%), the resonance shows a further red-shift. A subsequentpressure decrease of hydrogen (0 torr) drives the resonance back,yet again showing hysteresis effects. The net red-shift between thelast twomeasurements at 0 torr is around 3 nm.

To further corroborate the important role of nanofocusingin single-particle hydrogen sensing, we successively enlarge thedistance between the gold triangle and the palladium particle.Figure 4b presents the optical scattering spectra of the gold trianglesensor in response to hydrogen exposure, in which d is increased to70 nm. Just as in the case of d=10 nm, the sensor follows an overallsimilar trail on hydrogen absorption or desorption, but with muchsmaller spectral shifts. This is due to the fact that plasmon-inducedelectromagnetic fields decay exponentially away from the goldantenna surface. As a result, the LSPR shows a substantially reducedresponse to the dielectric change in a spot away from the nanofocus.Further evidence is presented in Fig. 4c, in which the distancebetween the gold triangle and the palladium particles is enlargedto d= 90 nm. In this case, the antenna resonance shows no spectralshift when the hydrogen pressure is 8 torr and exhibits a very smallred-shift at 16 torr. To effectively detect targeted events, the sensingelement should be placed as close as possible to the plasmonicnanofocus, inwhich the near-fields have their largest strength.

The shape of the plasmonic antenna is another crucialissue, which can significantly influence the hydrogen-sensor

performance. We demonstrate this aspect in Fig. 5, where we usea gold rod antenna. The length and width of the gold rod are130 nm and 80 nm, respectively, and its thickness is 40 nm. Thediameter and thickness of the palladium particle is 60 nm and40 nm, respectively. Its size is the same as that of the palladiumparticle next to the gold triangle antenna. Figure 5a shows theresults for a distance of d = 20 nm between the gold rod andthe palladium particle. It is evident that in general the responseof the gold rod sensor on hydrogen absorption or desorptionresembles that of the gold triangle sensor in Fig. 4a. Nevertheless,there are several dissimilarities that are noteworthy. First, theresonance does not show a significant shape change on hydrogenabsorption at 8 torr compared with 0 torr. Second, the magnitudesof red-shifts on hydrogen absorption and the hysteresis effects onhydrogen desorption are much smaller. The above observationscan be attributed to the fact that for the gold rod antenna theplasmon-induced near-fields are more located at the corners,whereas the sharp tip of the gold triangle antenna facilitates anexcellent nanfocus at the palladium particle spot, giving rise toa more sensitive detection response. When the distance betweenthe gold rod antenna and the palladium particle is enlarged tod = 50 nm, the overall response to hydrogen is smaller with respectto the case of d= 20 nm. Subsequently, when the distance is furtherincreased to d = 85 nm, the gold rod antenna can hardly sense theoptical changes of the palladium particle on hydrogen exposure.In essence, the triangle antenna sensor shows stronger feedback tohydrogen exposure than the rod antenna sensor as corroboratedby Figs 4a and 5a. The geometry of plasmonic antennas thus hasa substantial impact on determining the sensitivity of gas sensing atthe single-particle level.

As control experiments, we also measured the optical scatteringresponse of a single gold antenna and a single palladium nanoparti-cle on hydrogen exposure. As is evident from Fig. 6a, hydrogen doesnot influence the spectral position of the gold-antenna resonance.As Fig. 6b demonstrates, the palladium nanoparticles are nearlyindiscernible in a dark-field microscope owing to the extremelydamped plasmon excitations45. The scattering response of a singlepalladium nanoparticle shows unstructured and noisy spectra atdifferent hydrogen pressures (see Fig. 6b). In other words, a pal-ladium nanoparticle alone cannot fulfil the role of single-particlesensing in the visible range.





NATURE MATERIALS DOI: 10.1038/NMAT3029 ARTICLESGold antenna

Dark-field image

10 µm

Dark-field image

10 µm

33 torr

0 torr

33 torr

0 torr

Pd nanoparticle

550 600 650 700

Wavelength (nm)

500 550 600 650

Wavelength (nm)

Scat

tere

d in

tens

ity (

a.u.

)Sc

atte

red

inte

nsity

(a.

u.)

a

b

Figure 6 | Control experiments. a, Optical scattering measurements of asingle gold antenna. Upper right: Dark-field image of the gold antennas.Lower right: SEM image of a single gold antenna; scale bar 50 nm. Theplasmonic resonance of the gold antenna does not show any spectral shifton hydrogen exposure. b, Optical scattering measurements of a singlepalladium nanoparticle. Upper right: Dark-field image of the palladiumnanoparticles. Lower right: SEM image of a single palladium particle; scalebar 50 nm. The palladium nanoparticle alone shows an unstructured andnoisy spectrum. No useful information can be extracted onhydrogen exposure.

Resonant antenna-enhanced single-particle sensing pushes thesensitivity of plasmonic gas sensors to an ultimate limit and opensup myriad possibilities for detecting optically inactive (that is, notluminescent, with low oscillator strength or largely damped) speciesin a controlled fashion. The single-particle sensing strategy will haveprofound significance for the optical observation of chemical reac-tions and catalytic activities on a single platform in nanoreactors,and has the potential to be extended to biochemical systems inthe future. Moreover, antenna-enhanced sensing comprises a non-invasive and generalizable scheme that is applicable to a variety ofphysical and biochemical materials. For example, it can be directlyapplied to detect gases such as CO and NOx using platinum48,palladium or other inorganic catalysts, to detect CO2 adsorptionin metal–organic frameworks34 and to detect other gases usingappropriate analytes in the nanofocus. The device presented here isbiocompatible and can operate in aqueous environments. Ourworkcould pave the way towards the observation of catalytic reactionssuch as hydrogenation of carbohydrates59 on individual palladiumnanoparticles. The extension of indirect plasmonic sensing to theobservation of single chemical or biological events is challenging,but worth further investigation. Our sensing concept will alsostimulate the synthesis of advanced hybrid particles for plasmonicsensing. For example, it is possible to use appropriate biochemicallinkers to connect a gold particle with a Pd particle and to carry outhydrogen sensing at the single-particle level.We believe that this willbridge two important fields: plasmonics and biochemistry.

MethodsStructural fabrication. The antenna structures and alignment markers are firstdefined in poly(methyl methacrylate) resist using electron-beam lithography on aquartz substrate. The substrate is then covered with 2 nm chromium adhesion filmand 40 nm gold film using thermal evaporation followed by a lift-off procedure.Next, the substrate is coated once more with poly(methyl methacrylate) resist.Computer-controlled alignment at the sub-10-nm level using the gold alignmentmarkers is applied to ensure the accurate positioning of the single dot structure.Subsequently, the palladium dot is generated by covering the substrate with 2 nmchromium adhesion film and 40 nm palladium film using thermal evaporationfollowed by a lift-off procedure.

Experimental set-up. Ultrahigh-purity hydrogen (UHP 5.0) and nitrogen (UHP5.0) from Praxair were used with GFC 17 mass-flow controllers (Aalborg,Orangeburg, NY) to control the partial pressure of hydrogen in a stainless-steelcell. The total pressure was read by a NIST-certified diaphragm manometer gauge(no. 902346, Vacuum Research Corp., Pittsburgh, PA). The dark-field opticalset-up consists of a 100W tungsten lamp, a dark-field condenser and an objective.The scattered light was analysed by a silicon CCD (charge-coupled device)camera (1340×100 pixels) cooled by liquid nitrogen and attached to an imagingspectrometer. The detection took place normal to the sample surface throughthe objective optical axis (Zeiss dark-field objective, LC EC ‘Epiplan Neofluar’,50×, numerical aperture NA= 0.55), which is perpendicular to the illuminationpolarization. The illumination took place through the dark-field condenser. Themeasurements at different hydrogen concentrations were made after the systemwas in equilibrium in hydrogen–nitrogen atmosphere for 1 h. Each spectrumrepresents an average of 20 acquisitions that individually had an integration timeof 1 s at a single antenna structure. After a complete cycle, the system was purgedusing nitrogen overnight. The black curves imposed on the experimental spectra inFigs 4–6 were obtained using the Savitzky–Golay method by filtering the raw datawith a second-order polynomial and a frame size of 205.

Received 17 January 2011; accepted 18 April 2011; published online15 May 2011

References1. Schuller, J. A. et al. Plasmonics for extreme light concentration and

manipulation. Nature Mater. 9, 193–204 (2010).2. Stockman, M. I. Nanofocusing of optical energy in tapered plasmonic

waveguides. Phys. Rev. Lett. 93, 137404 (2004).3. Ropers, C. et al. Grating-coupling of surface plasmons onto metallic tips: A

nanoconfined light source. Nano Lett. 7, 2784–2788 (2007).4. Stockman, M. I., Faleev, S .V. & Bergman, D. J. Localization versus

delocalization of surface plasmons in nanosystems: Can one state have bothcharacteristics? Phys. Rev. Lett. 87, 167401 (2001).

5. Alu, A. & Engheta, N. Tuning the scattering response of optical nanoantennaswith nanocircuit loads. Nature Photon. 2, 307–310 (2008).

6. Shvets, G., Trendafilov, S., Pendry, J. B. & Sarychev, A. Guiding, focusing, andsensing on the subwavelength scale using metallic wire arrays. Phys. Rev. Lett.99, 053903 (2007).

7. Becker, J. et al. Plasmonic focusing reduces ensemble linewidth of silver-coatedgold nanorods. Nano Lett. 8, 1719–1723 (2008).

8. Lal, S., Link, S. & Halas, N. J. Nano-optics from sensing to waveguiding.Nature Photon. 1, 641–648 (2007).

9. Anker, J. N. et al. Biosensing with plasmonic nanosensors. Nature Mater. 7,442–453 (2008).

10. Liao, H. W., Nehl, C. L. & Hafner, J. H. Biomedical applications of plasmonresonant metal nanoparticles. Nanomedicine 1, 201–208 (2006).

11. Wang, F. & Shen, Y. R. General properties of local plasmons in metalnanostructures. Phys. Rev. Lett. 97, 206806 (2006).

12. Maier, S. A. Plasmonics: Fundamentals and Applications (Springer, 2007).13. Durach, M., Rusina, A. & Stockman, M. I. Toward full spatiotemporal control

on the nanoscale. Nano Lett. 7, 3145–3149 (2007).14. Kinkhabwala, A. et al. Large single-molecule fluorescence enhancements

produced by a bowtie nanoantenna. Nature Photon. 3, 654–657 (2009).15. Anger, P., Bharadwaj, P. & Novotny, L. Enhancement and quenching of

single-molecule fluorescence. Phys. Rev. Lett. 96, 113002 (2006).16. Kuhn, S., Hakanson, U., Rogobete, L. & Sandoghdar, V. Enhancement

of single-molecule fluorescence using a gold nanoparticle as an opticalnanoantenna. Phys. Rev. Lett. 97, 017402 (2006).

17. Li, J. F. et al. Shell-isolated nanoparticle-enhanced Raman spectroscopy.Nature 464, 392–395 (2010).

18. Camden, J .P. et al. Probing the structure of single-molecule surface-enhancedRaman scattering hot spots. J. Am. Chem. Soc. 130, 12616–12617 (2008).

19. Lim, D. K., Jeon, K. S., Kim, H. M., Nam, J. M. & Suh, Y. D.Nanogap-engineerable Raman-active nanodumbbells for single-moleculedetection. Nature Mater. 9, 60–67 (2010).

20. Taminiau, T. H., Stefani, F. D., Segerink, F. B. & Van Hulst, N. F. Opticalantennas direct single-molecule emission. Nature Photon. 2, 234–237 (2008).






21. Bharadwaj, P., Beams, R. & Novotny, L. Nanoscale spectroscopy with opticalantennas. Chem. Sci. 2, 136–140 (2011).

22. Schumacher, Th. et al. Nanoantenna-enhanced ultrafast nonlinear spectroscopyof a single gold nanoparticle. Preprint at http://arxiv.org/abs/1104.4855 (2011).

23. Sondergaard, T. et al. Extraordinary optical transmission enhanced bynanofocusing. Nano Lett. 10, 3123–3128 (2010).

24. Chen, C. J. & Osgood, R. M. Direct observation of the local-field-enhancedsurface photochemical-reactions. Phys. Rev. Lett. 50, 1705–1708 (1983).

25. Zhang, W. H., Huang, L. N., Santschi, C. & Martin, O. J. F. Trapping andsensing 10 nmmetal nanoparticles using plasmonic dipole antennas.Nano Lett.10, 1006–1011 (2010).

26. Aimovi, S. S., Kreuzer, M. P., González, M. U. & Quidant, R. Plasmonnear-field coupling in metal dimers as a step toward single-molecule sensing.ACS Nano 3, 1231–1237 (2009).

27. Adato, R. et al. Ultra-sensitive vibrational spectroscopy of proteinmonolayers with plasmonic nanoantenna arrays. Proc. Natl Acad. Sci.106, 19227–19232 (2009).

28. Hao, F., Nordlander, P., Sonnefraud, Y., Van Dorpe, P. & Maier, S. A.Tunability of subradiant dipolar and Fano-type plasmon resonances in metallicring/disk cavities: Implications for nanoscale optical sensing. ACS Nano 3,643–652 (2009).

29. Khalavka, Y., Becker, J. & Sonnichsen, C. Synthesis of rod-shaped goldnanorattles with improved plasmon sensitivity and catalytic activity. J. Am.Chem. Soc. 131, 1871–1875 (2009).

30. Reinhard, B. M., Sheikholeslami, S., Mastroianni, A., Alivisatos, A. P. &Liphardt, J. Use of plasmon coupling to reveal the dynamics of DNA bendingand cleavage by single EcoRV restriction enzymes. Proc. Natl Acad. Sci. 104,2667–2672 (2007).

31. Novo, C., Funston, A. M. & Mulvaney, P. Direct observation of chemicalreactions on single gold nanocrystals using surface plasmon spectroscopy.Nature Nanotech. 3, 598–602 (2008).

32. Aksu, S. et al. High-throughput nanofabrication of infrared plasmonicnanoantenna arrays for vibrational nanospectroscopy. Nano Lett. 10,2511–2518 (2010).

33. Bingham, J. M., Anker, J. N., Kreno, L. E. & Van Duyne, R. P. Gas sensingwith high-resolution localized surface plasmon resonance spectroscopy.J. Am. Chem. Soc. 132, 17358–17359 (2010).

34. Kreno, L. E., Hupp, J. T. & Van Duyne, R. P. Metal–organic framework thinfilm for enhanced localized surface plasmon resonance gas sensing.Anal. Chem.82, 8042–8046 (2010).

35. Lewis, F. A. The Palladium Hydrogen System (Academic Press, 1967).36. Vargas, W. E., Rojas, I., Azofeifa, D. E. & Clark, N. Optical and

electrical properties of hydrided palladium thin films studied by aninversion approach from transmittance measurements. Thin Solid Films 496,189–196 (2006).

37. Favier, F., Walter, E. C., Zach, M. P., Benter, T. & Penner, R. M. Hydrogensensors and switches from electrodeposited palladiummesowire arrays. Science293, 2227–2231 (2001).

38. Walter, E. C., Favier, F. & Penner, R. M. Palladium mesowire arrays forfast hydrogen sensors and hydrogen-actuated switches. Anal. Chem. 74,1546–1553 (2002).

39. Chadwick, B., Tann, J., Brungs, M. & Gal, M. A hydrogen sensor based on theoptical generation of surface plasmons in a palladium alloy. Sensors Actuators B17, 215–220 (1994).

40. Zeng, X. Q. et al. Hydrogen gas sensing with networks of ultrasmallpalladium nanowires formed on filtration membranes. Nano Lett. 11,262–268 (2010).

41. Zoric, I., Larsson, E. M., Kasemo, B. & Langhammer, C. Localizedsurface plasmons shed light on nanoscale metal hydrides. Adv. Mater.22, 4628–4633 (2010).

42. Bhuvana, T. & Kulkarni, G. U. A SERS-activated nanocrystalline Pdsubstrate and its nanopatterning leading to biochip fabrication. Small 4,670–676 (2008).

43. Johnson, J. L., Behnam, A., Pearton, S. J. & Ural, A. Hydrogen sensing usingPd-functionalized multi-layer graphene nanoribbon networks. Adv. Mater. 22,4877–4880 (2010).

44. Yang, F., Kung, S. C., Cheng, M., Hemminger, J. C. & Penner, R. M. Smaller isfaster and more sensitive: The effect of wire size on the detection of hydrogenby single palladium nanowires. ACS Nano 4, 5233–5244 (2010).

45. Langhammer, C., Yuan, Z., Zoric, I. & Kasemo, B. Plasmonic properties ofsupported Pt and Pd nanostructures. Nano Lett. 6, 833–838 (2006).

46. Langhammer, C., Zoric, I. & Kasemo, B. Hydrogen storage in Pd nanodiskscharacterized with a novel nanoplasmonic sensing scheme. Nano Lett. 7,3122–3127 (2007).

47. Pakizeh, T., Langhammer, C., Zoric, I., Apell, P. & Kall, M. IntrinsicFano interference of localized plasmons in Pd nanoparticles. Nano Lett. 9,882–886 (2009).

48. Larsson, E. M., Langhammer, C., Zoric, I. & Kasemo, B. Nanoplasmonicprobes of catalytic reactions. Science 326, 1091–1094 (2009).

49. Langhammer, C., Larsson, E. M., Kasemo, B.&Zoric, I. Indirect nanoplasmonicsensing: Ultrasensitive experimental platform for nanomaterials science andoptical nanocalorimetry. Nano Lett. 10, 3529–3538 (2010).

50. Langhammer, C., Zhdanov, V. P., Zoric, I. & Kasemo, B. Size-dependentkinetics of hydriding and dehydriding of Pd nanoparticles. Phys. Rev. Lett. 104,135502 (2010).

51. Abdelsalam, M. E., Mahajan, S., Bartlett, P. N., Baumberg, J. J. & Russell, A. E.SERS at structured palladium and platinum surfaces. J. Am. Chem. Soc. 129,7399–7406 (2007).

52. Wu, D. Y., Li, J. F., Ren, B. & Tian, Z. Q. Electrochemicalsurface-enhanced raman spectroscopy of nanostructures. Chem. Soc. Rev. 37,1025–1041 (2008).

53. Roy, R., Hohng, S. C. & Ha, T. A practical guide to single-molecule FRET.Nature Methods 5, 507–516 (2008).

54. Liu, N. et al. Three-dimensional photonic metamaterials at optical frequencies.Nature Mater. 7, 31–37 (2008).

55. Tian, Z. Q., Ren, B. & Wu, D. Y. Surface-enhanced Raman scattering: Fromnoble to transition metals and from rough surfaces to ordered nanostructures.J. Phys. Chem. B 106, 9463–9483 (2002).

56. Langhammer, C., Zhdanov, V. P., Zoric, I. & Kasemo, B. Size-dependenthysteresis in the formation and decomposition of hydride in metalnanoparticles. Chem. Phys. Lett. 488, 62–66 (2010).

57. Sachs, C. et al. Solubility of hydrogen in single-sized palladium clusters.Phys. Rev. B 64, 075408 (2001).

58. Hughes, R. C. & Schubert, W. K. Thin-films of Pd/Ni alloys for detection ofhigh hydrogen concentrations. J. Appl. Phys. 71, 542–544 (1992).

59. Trapp, O. et al. High-throughput kinetic study of hydrogenation overpalladium nanoparticles: Combination of reaction and analysis. Chem. Eur. J.14, 4657–4666 (2008).

AcknowledgementsWe would like to thank X. Meng for help with the metal deposition at the MicrolabFacility of the Electrical Engineering and Computer Science Department, University ofCalifornia, Berkeley. The SEM studies were supported by the Molecular Foundry at theNational Center for Electron Microscopy at Lawrence Berkeley National Laboratory. Theexperimental set-upwas funded by the grant ‘‘A Synergistic Approach to theDevelopmentof New Classes of Hydrogen Storage Materials’’ from the US Department of Energy,DE-AC03-76SF00098. We acknowledge S. Hein for his material visualizations. We thankTh. Schumacher and M. Lippitz for discussions and comments. We thank A. Tittl andN. Strohfeldt for help with the measurements and data analysis. N.L., M.L.T. and A.P.A.acknowledge financial support through the Plasmonic-Enhanced Catalysis Project ofthe Air Force Office of Science Research, award number FA9550-10-1-0504. M.H. andH.G. were financially supported by Deutsche Forschungsgemeinschaft (SPP1391 andFOR557), by BMBF (13N9048 and 13N10146) and by Landesstiftung BW.

Author contributionsAll authors contributed extensively to the work presented in this paper.

Additional informationThe authors declare no competing financial interests. Reprints and permissionsinformation is available online at http://www.nature.com/reprints. Correspondence andrequests for materials should be addressed to A.P.A.




http://arxiv.org/abs/1104.4855

http://www.nature.com/reprints


Documents

nmat3029