Near-field Optical Imaging of Carbon Nanotubes Achim Hartschuh, Huihong Qian Tobias Gokus, Department Chemie und Biochemie and CeNS, Universität München Neil Anderson, Lukas Novotny The Institute of Optics, University of Rochester, Rochester, NY let me start my thanking the organizers for inviting me to this excellent conference and for giving me the opportunity to report on our efforts on near-field optics The work I am going to talk about was done in roc, tueb, and siegen just recently we moved to munich Nice, 25.09.2006 Achim Hartschuh, Nano-Optics München

Why High Spatial Resolution Near-field Optics? Why Optics? electronic and vibronic energies DOS excited state dynamics.... why should we be interested in near-field optics on nanotubes that provides high spatial resolution? why optics? energy of light quanta is the region of the electronic and vibronic transitions If we have a straight, uniform and defect free nanotube, the optical properties will not vary along the nanotube and we do not need to have high spatial resolution Nice, 25.09.2006 Achim Hartschuh, Nano-Optics @ LMU

Achim Hartschuh, Nano-Optics @ LMU Why Near-field Optics? however, very often we have something like this, a kink Nice, 25.09.2006 Achim Hartschuh, Nano-Optics @ LMU

Achim Hartschuh, Nano-Optics @ LMU Why Near-field Optics? or even worse, a bent nanotube with particles, carbon sticking on the surface Nice, 25.09.2006 Achim Hartschuh, Nano-Optics @ LMU

Why Near-field Optics? spatial resolution is limited by diffraction l/2 or even worse, a bent nanotube with particles, carbon sticking on the surface Nice, 25.09.2006 Achim Hartschuh, Nano-Optics @ LMU

Achim Hartschuh, Nano-Optics @ LMU  Diffraction Limit Abbé, Arch. Mikrosk. Anat. 9, 413 (1873) Uncertainty relation: 1873 Abbe wrote down this formula telling us that two objects that are closer than dx can not be distinguished the formula can be easily derived from an uncertainty relation linking the uncertainty in space to the uncertainty in momentum in a microscope k varies with the focusing angle Nice, 25.09.2006 Achim Hartschuh, Nano-Optics @ LMU

Tip-Enhanced Spectroscopy Wessel, JOSA B 2, 1538 (1985) we take a sharp metal tip to interact with the objects independly Nice, 25.09.2006 Achim Hartschuh, Nano-Optics @ LMU

Tip-Enhanced Spectroscopy laser illuminated metal tip Novotny et al. PRL 79, 645 (1997) Theory: Enhanced electric field confined to tip apex Mechanism: Lightning rod and antenna effect, plasmon resonances Nice, 25.09.2006 Achim Hartschuh, Nano-Optics @ LMU

Tip-Enhanced Spectroscopy SEM micrograph diameter = 22 nm enhanced electric field confined within 20 nm ? ……………….Optical imaging with 20 nm resolution?! ……………….Signal enhancement !? Nice, 25.09.2006 Achim Hartschuh, Nano-Optics @ LMU

Experimental Setup + Confocal microscope Tip-sample distance control a sharp metal tip is held at constant height (~2nm) above the sample using a tuning-fork feedback mechanism Topography of the sample 2 nm K. Karrai et al., APL 66, 1842 (1995) Tip-sample distance control + Optical Images and Spectra Nice, 25.09.2006 Achim Hartschuh, Nano-Optics @ LMU

Tip-enhanced Spectroscopy (Examples…………) Raman scattering S.M. Stoeckle et al., Chem. Phys. Lett. 318, 131 (2000) N. Hayazawa et al., Chem. Phys. Lett. 335, 369 (2001) A. Hartschuh et al. Phys. Rev. Lett. 90, 95503 (2003) B. Pettinger et al., Phys. Rev. Lett. 92, 96101 (2004) N. Anderson et al., J. Am. Chem. Soc. 127, 2533 (2005) C.C. Neacsu et al., Phys. Rev. B 73, 193406 (2006) D. Methani, N. Lee, R. D. Hartschuh et al. J. Raman Spectrosc. 36, 1068 (2005) ................................. Two-photon excited fluorescence E.J. Sánchez et al., Phys. Rev. Lett. 82, 4014 (1999) H.G. Frey et al., Phys. Rev.Lett. 81, 5030 (2004) J. Farahani et al. Phys. Rev. Lett. 95, 017402 (2005) A. Hartschuh et al. Nano Lett. 5, 2310 (2005) Fluorescence Nice, 25.09.2006 Achim Hartschuh, Nano-Optics @ LMU

Single-Walled Carbon Nanotubes (SWCNT) Graphene sheet + roll up single-walled carbon nanotube diameter ~ 1-2 nm model length up to mm structure (n,m) determines properties: (n-m) mod 3 =0 metallic else semiconducting we can think of a cnt as being formed by a graphene sheet that has been rolled up to form a seemless zylinder from Shigeo Maruyama Nice, 25.09.2006 Achim Hartschuh, Nano-Optics @ LMU

Near-field Raman Imaging of SWCNTs Raman image (G’ band) Topography Hartschuh et al. PRL 90, 95503 (2003) 500 nm 500 nm this is one of the first near-field raman images we detected, about 3 years ago only SWCNT detected in optical image chemically specific optical contrast with 25 nm resolution enhanced field confined to tip Nice, 25.09.2006 Achim Hartschuh, Nano-Optics @ LMU

Near-field Raman Spectroscopy Raman image (G band) Topography image RBM at 199 cm-1 diam = 1.2 nm structure (n,m)(14,2) metallic SWCNT G RBM height: 0 - 1.9 nm Nice, 25.09.2006 Achim Hartschuh, Nano-Optics @ LMU

Resolution enhancement Farfield Near-field no tip same area with tip Nice, 25.09.2006 Achim Hartschuh, Nano-Optics @ LMU

Achim Hartschuh, Nano-Optics @ LMU Signal Enhancement Nice, 25.09.2006 Achim Hartschuh, Nano-Optics @ LMU

Achim Hartschuh, Nano-Optics @ LMU Signal Enhancement tip-enhanced signal > signal * 2500 Hartschuh et al. Phil. Trans. R. Soc. Lond A, 362 (2004) Nice, 25.09.2006 Achim Hartschuh, Nano-Optics @ LMU

Achim Hartschuh, Nano-Optics @ LMU Distance Dependence Enhanced Raman scattering signal ~ d-12 d development of new techniques and their applicaton tip-enhancement is near-field effect => tip has to be close to sample Nice, 25.09.2006 Achim Hartschuh, Nano-Optics @ LMU

Near-field Raman Spectroscopy N. Anderson, unpub. RBM G 30 nm steps wRBM=251 cm-1  (10,3) (semiconducting) wRBM=191 cm-1  (12,6) meanwhile the group at rochester is performing detailed near-field Raman studies the topo of a nanotube spectra taken in steps of about 30 nm initially show a rbm at indicating sc tube disappears and an rbm at 191 corresponding to a metallic nt appears we detect corresponding changes in g band in essence we believe to observe an imj (metallic) intramolecular junction (IMJ) pn-junction Nice, 25.09.2006 Achim Hartschuh, Nano-Optics @ LMU

Simultaneous Raman and PL Spectroscopy (7,5) (8,3) (9,1) Photoluminescence (6,4) Raman Emission Raman cnt are very particular material, at least for spectroscopists since we can observe both flu and Ra on the single object level with identical experimental conditions the fluorescence from cnt results from exc rec. and ist energy is directly determined by the structure, an 7,5 tube.... Excitation at 633 nm  Emission and Raman signals are spectrally isolated A. Hartschuh et al. Science 301, 1394 (2003) Nice, 25.09.2006 Achim Hartschuh, Nano-Optics @ LMU

Near-field optical imaging of SWCNTs Farfield Raman image Topography Raman (G-band) Photoluminescence  correlate Raman and photoluminescence properties  length of emissive segment ~ 70 nm changes in chirality (n,m)? coupling to substrate? DNA-wrapped SWCNTs on mica Nice, 25.09.2006 Achim Hartschuh, Nano-Optics @ LMU

Localized PL-Emisssion on Glass Photoluminescence Raman scattering SWCNTs on glass Emission spatially confined to within 10 – 20 nm  Localized excited states Bound excitons? role of defects? substrate? Photoluminescence if we deposite the raw nanotube material directly on glass, we detect extremely bright but highly confined PL signals Nice, 25.09.2006 Achim Hartschuh, Nano-Optics @ LMU

Near-field PL-Spectroscopy Topography Photoluminescence PL spectra: 30 nm steps between spectra 1 2 3 4 5 6 no SDS in the optical images.....  Emission energy can vary on the nanoscale (caused by changes in local dielectric environment e) A. Hartschuh et al. Nano Lett. 5, 2310 (2005) Nice, 25.09.2006 Achim Hartschuh, Nano-Optics @ LMU

Achim Hartschuh, Nano-Optics @ LMU Tip-enhanced Microscopy Spatial resolution < 15 nm  Signal amplification  Tip as nanoscale „light source“ whole story? Nice, 25.09.2006 Achim Hartschuh, Nano-Optics @ LMU

Achim Hartschuh, Nano-Optics @ LMU Signal Enhancement Raman scattering Photoluminescene Enhancement of incident field and scattered field SRaman ~ (Elocal / E0)2 (Elocal / E0) 2=f4 kex: enhanced excitation field SPL~ (Elocal / E0)2=f2 local field at tip field without tip Q is increased (Q0~10-4) cycling rate is increased krad: Purcell-effect krad + knonrad krad Q = the enhancement of the Raman signal results from the enhancement of the incident field and the scattered field and can be approximated by the fourth power of the field enhancement factor for fluorescence signal I think it is useful to consider the transitions rates involved the excitation rate is enhanced by the enhanced excitation field, this it the first part the tip will also modify the radiative rate because of a modified mode density. most likely it will be increased. an increase in the radiative rate corresponds to an increase in the Q and cycling rate and both will increase the detected signal however, the non-radiative rate will increase also because of energy transfer to the metal and quench the fluorescence kex krad knonrad Ein Eout PL Enhancement depends on Q0!  SPL~ f2 Q/Q0 dissipative energy transfer to metal quenching of PL knonrad: Nice, 25.09.2006 Achim Hartschuh, Nano-Optics @ LMU

Achim Hartschuh, Nano-Optics @ LMU Signal Enhancement Raman enhancement PL enhancement 500 nm 500 nm Far-field Raman ~ 2000 counts/s Near-field Raman ~ 4000 counts/s Raman-enhancement ~ 6000 / 2000 ~ 3 No far-field PL < 200 counts/s Near-field PL ~17000 counts/s PL-enhancement >17000 / 200 ~ 85 the enhancement of the Raman signal results from the enhancement of the incident field and the scattered field and can be approximated by the fourth power of the field enhancement factor for fluorescence signal I think it is useful to consider the transitions rates involved the excitation rate is enhanced by the enhanced excitation field, this it the first part the tip will also modify the radiative rate because of a modified mode density. most likely it will be increased. an increase in the radiative rate corresponds to an increase in the Q and cycling rate and both will increase the detected signal however, the non-radiative rate will increase also because of energy transfer to the metal and quench the fluorescence SRaman ~ f4 SPL~ f2 Q/Q0  PL quantum yield is increased by tip (SEF) Nice, 25.09.2006 Achim Hartschuh, Nano-Optics @ LMU

Achim Hartschuh, Nano-Optics @ LMU Signal Enhancement (First data) d knon-rad kex krad  PL quenching for very small distances  optimum distance for PL enhancement Nice, 25.09.2006 Achim Hartschuh, Nano-Optics @ LMU

Achim Hartschuh, Nano-Optics @ LMU Signal Enhancement (First data) P. Anger et al. d knon-rad kex krad Nice, 25.09.2006 Achim Hartschuh, Nano-Optics @ LMU

Achim Hartschuh, Nano-Optics @ LMU Signal Enhancement Topography Raman scattering Photoluminescence 290nm 290nm 290nm DNA-wrapped SWCNTs tip-nanotube-distance Raman scattering Photoluminescence (schematic) Nice, 25.09.2006 Achim Hartschuh, Nano-Optics @ LMU

Near-field Interactions Nanotubes: Uncertainty relation: Diffraction limit for propagating waves: High resolution provided by evanescent fields that have higher k-vectors: uncertainty relation for propagating light we have fixed |k|=2pi/lambda, diffraction limit tells us that we can not localize fields / intensities better than dx what does it mean for nts. the max wv has typical values on the order of 1 inverse nm, the states we can acces are limited to the photon cone here high resolution is provided by evanescent fields with higher k here: higher localization achieved with optical near-fields, fields are evanescent and can basically have arbitrary k the numbers we have are k_nf=0.2 nm-1 vs. k_BZ=1 nm-1 up to now no indication for other transitions or selection rules  k-vectors of tip enhanced fields extend through BZ !  selection rules for optical transitions? Nice, 25.09.2006 Achim Hartschuh, Nano-Optics @ LMU

Achim Hartschuh, Nano-Optics @ LMU Summary High-resolution optical microscopy of carbon nanotubes using a sharp laser-illuminated metal tip PL and Raman spectroscoy and imaging Spatial resolution < 15 nm Signal enhancement Results Resolved RBM variations (IMJ) on the nanoscale Non-uniform emission energies that result from local variations of dielectric environment Strongly confined emission signals  bound excitons? PL-Quantum yield can be enhanced by metal tip strong variation consequence of sample material, there is certainly better defined material for which no such variations occur, on the other hand, interaction with the local environment also essential for perfect nanotube Nice, 25.09.2006 Achim Hartschuh, Nano-Optics @ LMU

Achim Hartschuh, Nano-Optics München Thanks to PC Tuebingen: Alfred J. Meixner Mathias Steiner Antonio Virgilio Failla PC Siegen: Gregor Schulte Funding: DFG, Cm Siegen, NSF, FCI Nice, 25.09.2006 Achim Hartschuh, Nano-Optics München

Achim Hartschuh, Nano-Optics @ LMU Acknowledgement PC Tuebingen: Alfred J. Meixner Mathias Steiner Antonio Virgilio Failla PC Siegen: Gregor Schulte Funding: DFG, Cm Siegen, NSF, FCI Nice, 25.09.2006 Achim Hartschuh, Nano-Optics @ LMU