- Email: [email protected]
Extraordinary tuning of a nanocavity by a near-field probe
Available online at www.sciencedirect.com
Photonics and Nanostructures – Fundamentals and Applications 9 (2011) 269–275 www.elsevier.com/locate/photonics
Extraordinary tuning of a nanocavity by a near-field probe Benoit Cluzel a, Loı¨c Lalouat a, Philippe Velha b,c,d,*, Emmanuel Picard b, Emmanuel Hadji b, David Peyrade d, Fre´de´rique de Fornel a a Groupe d’Optique de Champ Proche, Laboratoire Interdisciplinaire Carnot de Bourgogne, UMR CNRS 5209, LRC SiNOPTIC CEA n8DSM-08-36, Universite´ de Bourgogne, 9, Avenue A. Savary, BP 47870, F-21078 Dijon, France b SiNaPS-MINATEC, CEA Grenoble, 17 rue des Martyrs, F-38054 Grenoble, France c Laboratoire Charles Fabry de l’Institut d’Optique, CNRS, Universite´ de Paris-Sud, Campus Polytechnique, RD 128, F-91127 Palaiseau, France d Laboratoire des Technologies de la Microe´lectronique, CNRS, 17 rue des Martyrs, F-38054 Grenoble, France
Received 15 November 2010; received in revised form 28 April 2011; accepted 2 May 2011 Available online 13 May 2011
Abstract We report here an experimental observation of an extraordinary near-field interaction between a local probe and a small-volume solid-state nanocavity. We directly compare the normally observed near-field interaction regime driven by the perturbation theory and then report the extraordinary interaction regime. Subsequently, we show that the cavity can take up to 2 min to recover from this interaction after removing the probe and that leads to an extraordinary blue-shift of the cavity resonance wavelength (15 nm) which depends on the probe motion above the cavity and not the position. The reasons for this effect are not fully understood yet but we try to give some explanations. Crown Copyright # 2011 Published by Elsevier B.V. All rights reserved. Keywords: Photonic crystals; Near-field optics; Optomechanics; Silicon photonics; Microcavity; Anomalous regime; Nanotechnology; Tuning; Extraordinary regime
1. Introduction 1.1. General introduction Near-field microscopy techniques are devoted to visualize and analyse matter properties at the nanoscale. Among the near-field techniques that have been studied since the early 1980s, the optical near-field techniques which allow probing the light at a subwavelength scale, have been intensively investigated with the emerging field of nanophotonics [1,2]. In this
* Corresponding author at: SiNaPS-MINATEC, CEA Grenoble, 17 rue des Martyrs, F-38054 Grenoble, France. Tel.: +44 0141 330 6022; fax: +44 0141 330 4907. E-mail address: [email protected] (P. Velha).
context, scanning near-field optical microscopy (SNOM) techniques have proven their ability to analyse and visualize the spatial [3], spectral [4] and temporal [5] light behaviour in integrated optics components such as photonic crystals or plasmonic devices. However, in the same way that the scanning tunnelling microscopy (STM) and atomic force microscopy (AFM) probes permit the manipulation of individual atoms [6,7], molecules [8] or particles [9], the SNOM probes can also be used to manipulate the properties of confined electromagnetic fields [10–12]. The near-field interaction involved in these recent experiments relies on a local perturbation of the electromagnetic field induced by the probe. In this letter, we report an experimental observation of an extraordinary near-field interaction that occurs between a local probe and a small-volume solid-state
1569-4410/$ – see front matter. Crown Copyright # 2011 Published by Elsevier B.V. All rights reserved. doi:10.1016/j.photonics.2011.05.002
270
B. Cluzel et al. / Photonics and Nanostructures – Fundamentals and Applications 9 (2011) 269–275
cavity properties leads to an extraordinary blue-shift of the cavity resonance wavelength (15 nm) and depends on the probe motion inside the cavity optical near-field. The nano-cavities studied in this work are similar to that one reported in our previous [13,14] work. As shown in the SEM view of Fig. 1a, they consists in a Fabry Perot-like resonator composed of two mirrors etched in a silicon ridge waveguide and designed to suppress the radiation losses at the mirror termination [15]. For a Transverse Electrical polarization (TE with the H-field taken as normal to the substrate plane) the nanocavity exhibits a single resonance at telecommunication wavelength [16]. A typical high resolution transmittance spectrum of a nanocavity is given in Fig. 2. 2. Experimental Fig. 1. (a) Scanning electron microscope view of the studied nanocavity. Air holes are nano-patterned inside a silicon ridge waveguide. (b) Three-dimensional calculation of the electric field distribution of the nanocavity at resonance. We define the axis as X being perpendicular to waveguide direction and as Y the axis along the waveguide.
nanocavity that cannot be explained by the standard perturbation theory. In the first part of this letter, we directly compare the well-known near-field interaction regime driven by the perturbation theory and the reported extraordinary interaction regime. Secondly, we show that the interaction leads to a modification of the intrinsic properties of the cavity, which can persist up to 2 min after removing the probe from the cavity optical near-field. Finally, we show that the modification of the
As reported in our previous works [13,14] and confirmed by other groups [11,12] introducing a nearfield probe (dielectric: SiO2, Silicon and other related materials or non-magnetic metals: Au, Al) inside the electromagnetic field induces a local adiabatic perturbation of the effective index. This perturbation strength is proportional to electrical field distribution within the resonator. Consequently, scanning the near-field probes inside the resonator optical near-field while recording the far-field resonator transmittance as a function of the probe position, gives a map of the electrical field distribution of the resonator eigenmode. A typical nearfield image recorded above the nanocavity in such an interaction scanning mode is plotted in Fig. 3a. As a comparison, we also show in Fig. 1b the electric field distribution above the cavity at resonance calculated by using a 3D modal method [17]). 3. Results 3.1. Experimental anomaly
Fig. 2. High resolution transmittance spectrum of the nanocavity at resonance. The Lorentzian-shaped cavity resonance is modulated by high frequency oscillations corresponding to the interferences of the light bouncing between the two cleaved facets of the sample.
However, in several cases we measured cavities which do not follow the perturbation theory and instead exhibit an extraordinary regime of perturbation. At first, this regime seems to occur for random structures. No micro-structural modification between devices was found under SEM observation. At last, we found that the phenomenon reported hereafter seems to be correlated with the presence of surface bonds. It was systematically observed that soaking the structures in a 1% diluted hydrofluorhydric acid solution suppress the reported ‘‘extraordinary’’ interaction regime.
B. Cluzel et al. / Photonics and Nanostructures – Fundamentals and Applications 9 (2011) 269–275
271
Fig. 3. (a) Optical near-field image of the nanocavity resonance recorded in interaction scanning mode and for an interaction between the probe and the cavity driven by the perturbation theory. (b) Optical near-field image of the nanocavity resonance recorded in interaction scanning mode and for the reported extraordinary interaction between the probe and the cavity. On the two images the exact cavity centre is pointed by the arrow.
Nevertheless over several months the ‘‘extraordinary’’ regime has been observed reproductively on several cavities with various Q-factor (Q = 1000, 3000, 7000, 10,000, 20,000 and 40,000) and by using metallic (Au-coated probes) as well as dielectric (SiO2) probes. For the sake of brevity, we only present the result obtained on the same cavity (Q = 3000) with the same near-field probe (a chemically etched silica optical fibre). All the results shown can be transposed to all the other cavities with highest or lowest Q-factor we have investigated.
3.2. Differences between ‘‘standard’’ and ‘‘extraordinary’’ regime Fig. 4a presents a computed (3D modal method [17]) electrical field distribution taken across the waveguide (direction X see Fig. 1) on the central part of the cavity. In the case of a cavity which exhibits a ‘‘standard’’ interaction regime, according to the perturbation theory (Fig. 4b), the perturbed cavity transmittance which is inversely proportional the field distribution exhibits features comparable to the computed field distribution. In the case of the ‘‘extraordinary’’ interaction for two slightly different scans over the cavity though (Fig. 4c and d), the transmittance starts to resemble to the inverse of the field distribution but, as soon as the probe is removed from the cavity optical near-field, the cavity transmittance takes a few seconds to reach its initial value (Tmax). For a given cavity, we found that such persistence is systematically observed in the case of the reported ‘‘extraordinary’’ interaction.
3.3. Time constant characterisation We investigated then the time constants that are involved in the reported interaction, for this purpose, the transmission of the cavity is measured at resonance. The probe is first approached vertically from the cavity centre and stays ‘‘motionless’’ during a given waiting time (tw). Next, we remove the probe in a few ms from the cavity optical near-field and record the cavity transmittance temporal evolution. The schematic of this experiment is depicted in Fig. 5a. We repeated the experiment for several tw ranging from few seconds to few minutes as shown in Fig. 5b. Clearly, the cavity transmittance does not come back instantaneously to its initial value as predicted by the perturbation theory. As tw increases from 10 to 120 s, the time required to reach a transmittance value of Tmax/2 increases non-linearly from 4.5 s to 87 s. For tw over 120 s, the temporal evolution stays comparable to the one recorded at tw = 120 s, suggesting that an equilibrium has been achieved. 3.4. Follow up of the cavity dynamic While repeating the experiment we recorded the entire cavity transmittance spectra. The cavity resonance wavelength was found to be blue-shifted in total disagreement with the perturbation regime [10] which predicts a red-shift. Moreover, moving the probe above the cavity modifies the observed blue-shift value. Therefore, in order to quantify this resonance shift, we next recorded continuously the cavity transmittance spectra during a controlled probe scan above the whole structure. Preliminary, the probe was positioned at
272
B. Cluzel et al. / Photonics and Nanostructures – Fundamentals and Applications 9 (2011) 269–275
Fig. 4. (a) Cross-section of the electrical field intensity within the cavity at the resonance wavelength. (b) Cross-section of the standard perturbation of the cavity by the probe (according to the perturbation theory). (c, d) Different cross-sections of the extraordinary perturbation induced by the probe motion. The topographical lines corresponding to the probe motions associated to the optical measurements of (b–d) are plotted on the bottom.
1.5 mm away from the cavity centre and its motion was adjusted to be parallel to the waveguide axis. Then, as depicted in Fig. 6a, from t = 0 s, the probe is progressively approached from the cavity at a 0.2 Hz longitudinal scanning rate (around 1 mm/s). At t = t1, i.e. as soon as the probe scans the cavity centre, we set the probe motion to a single line scan along the cavity axis. This probe motion is maintained during a given time and at t = t2 the probe is instantaneously removed into the far-field. During all this procedure, the transmittance spectra were recorded continuously at a rate of 0.3 Hz. The transmittance spectrum evolution as a function of the probe scan is presented on a movie available
online and is summarized in Fig. 6c while the cavity resonance wavelength evolution extracted from the movie is plotted in Fig. 6b. All the recorded spectra exhibit a single resonance with high frequency oscillations due to Fabry Perot interferences between the ridge waveguide cleaved facets. As long as the probe is scanned out of the optical near-field of the nanocavity, the resonance stays unchanged. As soon as the probe penetrates the optical near-field of the cavity, the resonance wavelength starts blue-shifting. For t < t1, the cavity resonance decreases from 1585.4 nm to 1576 nm. Next, for t > t1, when the probe scan is blocked along a single axis centred on the cavity, the resonance wavelength continuously decreases to reach a
B. Cluzel et al. / Photonics and Nanostructures – Fundamentals and Applications 9 (2011) 269–275
273
Fig. 5. (a) Schematic view of the experiment: the near-field probe is maintained in the optical near-field of the cavity for a given waiting time tw. At t = 0 s, the probe is removed instantaneously into the far-field while recording for t > 0 s, the cavity transmittance temporal evolution. (b) Experimental measurements recorded for different waiting times: tw = 10, 30, 60 and 120 s.
final equilibrium value of 1569.4 nm at t1 + 300 s. For longer time, as long as the probe scan is maintained above the cavity, this resonance wavelength stays constant. Finally, we removed instantaneously the probe from the cavity optical near-field at t = t2 and clearly observed that the resonance wavelength increased progressively to reach its initial value at t2 + 150 s. At last, since the cavity peak transmittance as well as its resonance wavelength, come back to their initial values at the end of the experiment, the cavity was not damaged during the reported experiments.
3.5. Summary of the observations First, we found that the probe cavity interaction results in an extraordinary blue-shift of the cavity resonance wavelength, instead of a red-shift as expected from perturbation theory [13,14]. Next, the maximal blue-shift value is measured to be dl/l 1% (dl = 15 nm), an ‘‘extraordinary’’ value to be compared to the red-shift values of dl/l 0.01% (dl 0.2 nm) usually achieved with a similar probe. In addition, we find that a dynamical equilibrium state, which depends on the probe motion above the cavity, can be obtained. At last, the equilibrium state needs several tens of seconds to be achieved, which are
extraordinary long for an interaction relying only on a pure electromagnetic phenomenon. 4. Possible explanations In order to explain these unexpected results, one may consider several hypotheses. Assuming that the cavity length is not squeezed during the experiment the resonance shift is either due to a relative change of the refractive index dn/n 1% or to a change in the filling of the holes. Such refractive index change can be associated to either an electro-optic or a thermo-optic effect, which are the two dominant effects in silicon. We believe that the electro-optic effect is not involved for two strong reasons: - the time constant for the plasma dispersion effect in silicon is typically a few ns. - The density of carriers needed for such index change is in the excess of 1019 cm 3. Therefore, with such long time constants, thermooptical effects are more likely to be involved. However, in light of the temperature dependence of silicon refractive index [18–20], the relative temperature change corresponding to the reported dn/n 1% is higher than 50%, which corresponds to a cavity cooling
274
B. Cluzel et al. / Photonics and Nanostructures – Fundamentals and Applications 9 (2011) 269–275
Fig. 6. (a) Schematic view of the experiment. The cavity transmittance spectra are recorded during a probe scan: for t = 0 to t1, the probe is approached from the cavity, for t1 < t < t2, the probe motion is blocked to a single axis scan centred on the cavity. At t = t2, the probe is removed into the far field. (b) Temporal evolution of the cavity resonance wavelength extracted from the transmittance spectra recorded during the probe scan. The corresponding movie is available online. (c) Typical transmittance spectrum snapshots at different time during the experiment.
from an initial temperature of 300 K to a temperature lower than 150 K. Another possible explanation compatible with the wavelength shift amplitude may be the presence of trapped water inside the holes of the mirrors of the nanocavity which are affected by the motion of the near field probe on the top of the cavity. Even if we cannot be conclusive on this explanation it seems the most attractive. Firstly, the silica near-field probes are well-known to be hydrophilic and scanning the probe above holes filled with water could result in water being disturbed by the probe. Next, an FDTD simulation of a cavity, assuming that all the holes of the mirror are filled with water (n = 1.3) shows a 19-nm red shift of the resonance wavelength, in comparison to the same cavity without water. Then, measuring experimentally the transmit-
tance of an immersed nanocavity which does not exhibit the extraordinary shift leads to the observation of a 13 nm red-shift of its resonance wavelength. Those 3 values are all of the same order but the discrepancies are difficult to explain. Other works [21–23] have also shown that infiltration of liquids in holes can be used to tune photonic structures. However, these infiltration experiments have a time response of less than a second which lacks to explain our response time which can be up to 2 min. In addition, the experiment where the tip stays motionless on the centre of the cavity (Fig. 5) is hardly explained with this nanofluidics approach even if the few nanometre lateral oscillations of the probe due to the shear-force back could lead to weak amplitude water displacement at the surface of the nanocavity. Finally, the fact that the ‘‘extraordinary’’ regime is suppressed by a diluted HF etching (10 s in 1% HF
B. Cluzel et al. / Photonics and Nanostructures – Fundamentals and Applications 9 (2011) 269–275
solution) of the natural oxidization layer of silicon surface implies that this surface layer is deeply related to the reported phenomenon. Thus, combined to the fact that that the cavities are produced by dry etching without any wet etching post processing like HF, the presence of surface bonds which can trap water could explained the presence of water on the surface. Again another interpretation could rely on the presence of trapped electrostatic surface charges which could be created by the friction of the near-field probe on the surface. The presence of such charges could lead the existence of a strong electrostatic field able to strain the cavity by electrostriction. In this case, the dynamics of trapping and release of such charges [24] could agree with the time scale of the reported phenomenon. Against this last possibility remains the fact that the optical mode needs to be present for the effect to take place. Independently of the real causes, such an experiment shows that the preparation of a sample’s surface could have a tremendous influence in the field of NOEMS and nanophotonics. 5. Conclusions In conclusion, we reported a near-field interaction between a nanometric probe and a small-volume nanocavity, which allows the dynamical tuning of the cavity resonance over an extraordinarily broad spectral range. In light of all the experimental results, we were not able to identify rigorously the fundamental reasons of the observed interaction. Pure electro-thermo-optics effects as well as nanofluidic interactions with the nearfield probe do not seem to fully and satisfactorily explain all the experimental results. The most important conclusion of this experiment is that the surface preparation can have a drastic effect on NOEMS and consequently, we strongly believe that a cross-disciplinary analysis is required for further investigations of the reported phenomenon. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/ j.photonics.2011.05.002.
275
References [1] F. de Fornel, Evanescent Waves from Newtonian to Atomic Optics, Springer Verlag, Heidelberg, 2001. [2] L. Novotny, B. Hecht, Principles of Nano-Optics, Cambridge University Press, 2006. [3] P.T. Rakich, M.S. Dahlem, S. Tandon, M. Ibanescu, M. Soljacic, G.S. Petrtich, J.D. Joannopulos, L.A. Kolodziejski, E.P. Ippen, Nat. Mater. 5 (2006) 93. [4] N. Louvion, D. Ge´rard, J. Mouette, F. de Fornel, C. Seassal, X. Letartre, A. Rahmani, S. Callard, Phys. Rev. Lett. 94 (2005) 113907. [5] M.L.M. Balistreri, H. Gersen, J.P. Korterik, L. Kuipers, N.F. van Hulst, Science 294 (2001) 1080. [6] D.M. Eigler, E.K. Schweizer, Nature 344 (1990) 524. [7] M.F. Crommie, C.P. Lutz, D.M. Eigler, Science 262 (1993) 218. [8] R.D. Piner, et al. Science 283 (1999) 661. [9] T.R. Ramachandran, C. Baur, A. Bugacov, A. Madhukar, B.E. Koel, A. Requicha, G. Gazen, Nanotechnology 9 (1998) 237. [10] L. Lalouat, B. Cluzel, P. Velha, E. Picard, D. Peyrade, J.P. Hugonin, P. Lalanne, E. Hadji, F. de Fornel, Phys. Rev. B 76 (2007) 041102. [11] A.F. Koenderink, M. Kafesaki, B.C. Buchler, V. Sandoghdar, Phys. Rev. Lett. 95 (2005) 153904. [12] S. Mujumdar, A.F. Koenderink, T. Su¨nner, B.C. Buchler, M. Kamp, A. Forchel, V. Sandoghdar, Opt. Express 15 (2007) 17214. [13] B. Cluzel, L. Lalouat, P. Velha, E. Picard, D. Peyrade, J.C. Rodier, T. Charvolin, P. Lalanne, F. de Fornel, E. Hadji, Opt. Express 16 (2008) 279. [14] L. Lalouat, B. Cluzel, F. de Fornel, P. Velha, P. Lalanne, D. Peyrade, E. Picard, T. Charvolin, E. Hadji, App. Phys. Lett. 92 (2008) 111111. [15] C. Sauvan, G. Lecamp, P. Lalanne, J.P. Hugonin, Opt. Express 13 (2005) 245. [16] P. Velha, E. Picard, T. Charvolin, E. Hadji, J.C. Rodier, P. Lalanne, D. Peyrade, Opt. Express 15 (2007) 16090. [17] E. Silberstein, P. Lalanne, J.P. Hugonin, Q. Cao, J. Opt. Soc. Am. A 18 (2001) 2865. [18] J.A. McCaulley, V.M. Donnelly, M. Vernon, I. Taha, Phys. Rev. B 49 (1994) 7408. [19] B.J. Frey, D.B. Leviton, T.J. Madison, Proc. SPIE 7273 (2006). [20] M.W. Pruessner, T.H. Stievater, M.S. Ferraro, W.S. Rabinovich, Opt. Express 15 (2007) 7757. [21] S. Vignolini, F. Riboli, F. Intonti, M. Belotti, M. Gurioli, Y. Chen, M. Colocci, L.C. Andreani, W.S. Diederik, Phys. Rev. E 78 (2008) 45603. [22] M. Ebnali-Heidari, C. Grillet, C. Monat, B.J. Eggleton, Opt. Express 17 (2009) 1628. [23] C.L. Smith, U. Bog, S. Tomljenovic, M.W. Lee, D.K. Wu, L. O’Faolain, C. Monat, C. Grillet, T.F. Krauss, C. Karnutsch, R.C. McPhedran, B.J. Eggleton, Opt. Express 16 (2008) 15887. [24] J.S. Milne, A.C.H. Rowe, S. Arscott, Ch. Renner, Phys. Rev. Lett. 105 (22) (2010) 226802.
Photonics and Nanostructures – Fundamentals and Applications 9 (2011) 269–275 www.elsevier.com/locate/photonics
Extraordinary tuning of a nanocavity by a near-field probe Benoit Cluzel a, Loı¨c Lalouat a, Philippe Velha b,c,d,*, Emmanuel Picard b, Emmanuel Hadji b, David Peyrade d, Fre´de´rique de Fornel a a Groupe d’Optique de Champ Proche, Laboratoire Interdisciplinaire Carnot de Bourgogne, UMR CNRS 5209, LRC SiNOPTIC CEA n8DSM-08-36, Universite´ de Bourgogne, 9, Avenue A. Savary, BP 47870, F-21078 Dijon, France b SiNaPS-MINATEC, CEA Grenoble, 17 rue des Martyrs, F-38054 Grenoble, France c Laboratoire Charles Fabry de l’Institut d’Optique, CNRS, Universite´ de Paris-Sud, Campus Polytechnique, RD 128, F-91127 Palaiseau, France d Laboratoire des Technologies de la Microe´lectronique, CNRS, 17 rue des Martyrs, F-38054 Grenoble, France
Received 15 November 2010; received in revised form 28 April 2011; accepted 2 May 2011 Available online 13 May 2011
Abstract We report here an experimental observation of an extraordinary near-field interaction between a local probe and a small-volume solid-state nanocavity. We directly compare the normally observed near-field interaction regime driven by the perturbation theory and then report the extraordinary interaction regime. Subsequently, we show that the cavity can take up to 2 min to recover from this interaction after removing the probe and that leads to an extraordinary blue-shift of the cavity resonance wavelength (15 nm) which depends on the probe motion above the cavity and not the position. The reasons for this effect are not fully understood yet but we try to give some explanations. Crown Copyright # 2011 Published by Elsevier B.V. All rights reserved. Keywords: Photonic crystals; Near-field optics; Optomechanics; Silicon photonics; Microcavity; Anomalous regime; Nanotechnology; Tuning; Extraordinary regime
1. Introduction 1.1. General introduction Near-field microscopy techniques are devoted to visualize and analyse matter properties at the nanoscale. Among the near-field techniques that have been studied since the early 1980s, the optical near-field techniques which allow probing the light at a subwavelength scale, have been intensively investigated with the emerging field of nanophotonics [1,2]. In this
* Corresponding author at: SiNaPS-MINATEC, CEA Grenoble, 17 rue des Martyrs, F-38054 Grenoble, France. Tel.: +44 0141 330 6022; fax: +44 0141 330 4907. E-mail address: [email protected] (P. Velha).
context, scanning near-field optical microscopy (SNOM) techniques have proven their ability to analyse and visualize the spatial [3], spectral [4] and temporal [5] light behaviour in integrated optics components such as photonic crystals or plasmonic devices. However, in the same way that the scanning tunnelling microscopy (STM) and atomic force microscopy (AFM) probes permit the manipulation of individual atoms [6,7], molecules [8] or particles [9], the SNOM probes can also be used to manipulate the properties of confined electromagnetic fields [10–12]. The near-field interaction involved in these recent experiments relies on a local perturbation of the electromagnetic field induced by the probe. In this letter, we report an experimental observation of an extraordinary near-field interaction that occurs between a local probe and a small-volume solid-state
1569-4410/$ – see front matter. Crown Copyright # 2011 Published by Elsevier B.V. All rights reserved. doi:10.1016/j.photonics.2011.05.002
270
B. Cluzel et al. / Photonics and Nanostructures – Fundamentals and Applications 9 (2011) 269–275
cavity properties leads to an extraordinary blue-shift of the cavity resonance wavelength (15 nm) and depends on the probe motion inside the cavity optical near-field. The nano-cavities studied in this work are similar to that one reported in our previous [13,14] work. As shown in the SEM view of Fig. 1a, they consists in a Fabry Perot-like resonator composed of two mirrors etched in a silicon ridge waveguide and designed to suppress the radiation losses at the mirror termination [15]. For a Transverse Electrical polarization (TE with the H-field taken as normal to the substrate plane) the nanocavity exhibits a single resonance at telecommunication wavelength [16]. A typical high resolution transmittance spectrum of a nanocavity is given in Fig. 2. 2. Experimental Fig. 1. (a) Scanning electron microscope view of the studied nanocavity. Air holes are nano-patterned inside a silicon ridge waveguide. (b) Three-dimensional calculation of the electric field distribution of the nanocavity at resonance. We define the axis as X being perpendicular to waveguide direction and as Y the axis along the waveguide.
nanocavity that cannot be explained by the standard perturbation theory. In the first part of this letter, we directly compare the well-known near-field interaction regime driven by the perturbation theory and the reported extraordinary interaction regime. Secondly, we show that the interaction leads to a modification of the intrinsic properties of the cavity, which can persist up to 2 min after removing the probe from the cavity optical near-field. Finally, we show that the modification of the
As reported in our previous works [13,14] and confirmed by other groups [11,12] introducing a nearfield probe (dielectric: SiO2, Silicon and other related materials or non-magnetic metals: Au, Al) inside the electromagnetic field induces a local adiabatic perturbation of the effective index. This perturbation strength is proportional to electrical field distribution within the resonator. Consequently, scanning the near-field probes inside the resonator optical near-field while recording the far-field resonator transmittance as a function of the probe position, gives a map of the electrical field distribution of the resonator eigenmode. A typical nearfield image recorded above the nanocavity in such an interaction scanning mode is plotted in Fig. 3a. As a comparison, we also show in Fig. 1b the electric field distribution above the cavity at resonance calculated by using a 3D modal method [17]). 3. Results 3.1. Experimental anomaly
Fig. 2. High resolution transmittance spectrum of the nanocavity at resonance. The Lorentzian-shaped cavity resonance is modulated by high frequency oscillations corresponding to the interferences of the light bouncing between the two cleaved facets of the sample.
However, in several cases we measured cavities which do not follow the perturbation theory and instead exhibit an extraordinary regime of perturbation. At first, this regime seems to occur for random structures. No micro-structural modification between devices was found under SEM observation. At last, we found that the phenomenon reported hereafter seems to be correlated with the presence of surface bonds. It was systematically observed that soaking the structures in a 1% diluted hydrofluorhydric acid solution suppress the reported ‘‘extraordinary’’ interaction regime.
B. Cluzel et al. / Photonics and Nanostructures – Fundamentals and Applications 9 (2011) 269–275
271
Fig. 3. (a) Optical near-field image of the nanocavity resonance recorded in interaction scanning mode and for an interaction between the probe and the cavity driven by the perturbation theory. (b) Optical near-field image of the nanocavity resonance recorded in interaction scanning mode and for the reported extraordinary interaction between the probe and the cavity. On the two images the exact cavity centre is pointed by the arrow.
Nevertheless over several months the ‘‘extraordinary’’ regime has been observed reproductively on several cavities with various Q-factor (Q = 1000, 3000, 7000, 10,000, 20,000 and 40,000) and by using metallic (Au-coated probes) as well as dielectric (SiO2) probes. For the sake of brevity, we only present the result obtained on the same cavity (Q = 3000) with the same near-field probe (a chemically etched silica optical fibre). All the results shown can be transposed to all the other cavities with highest or lowest Q-factor we have investigated.
3.2. Differences between ‘‘standard’’ and ‘‘extraordinary’’ regime Fig. 4a presents a computed (3D modal method [17]) electrical field distribution taken across the waveguide (direction X see Fig. 1) on the central part of the cavity. In the case of a cavity which exhibits a ‘‘standard’’ interaction regime, according to the perturbation theory (Fig. 4b), the perturbed cavity transmittance which is inversely proportional the field distribution exhibits features comparable to the computed field distribution. In the case of the ‘‘extraordinary’’ interaction for two slightly different scans over the cavity though (Fig. 4c and d), the transmittance starts to resemble to the inverse of the field distribution but, as soon as the probe is removed from the cavity optical near-field, the cavity transmittance takes a few seconds to reach its initial value (Tmax). For a given cavity, we found that such persistence is systematically observed in the case of the reported ‘‘extraordinary’’ interaction.
3.3. Time constant characterisation We investigated then the time constants that are involved in the reported interaction, for this purpose, the transmission of the cavity is measured at resonance. The probe is first approached vertically from the cavity centre and stays ‘‘motionless’’ during a given waiting time (tw). Next, we remove the probe in a few ms from the cavity optical near-field and record the cavity transmittance temporal evolution. The schematic of this experiment is depicted in Fig. 5a. We repeated the experiment for several tw ranging from few seconds to few minutes as shown in Fig. 5b. Clearly, the cavity transmittance does not come back instantaneously to its initial value as predicted by the perturbation theory. As tw increases from 10 to 120 s, the time required to reach a transmittance value of Tmax/2 increases non-linearly from 4.5 s to 87 s. For tw over 120 s, the temporal evolution stays comparable to the one recorded at tw = 120 s, suggesting that an equilibrium has been achieved. 3.4. Follow up of the cavity dynamic While repeating the experiment we recorded the entire cavity transmittance spectra. The cavity resonance wavelength was found to be blue-shifted in total disagreement with the perturbation regime [10] which predicts a red-shift. Moreover, moving the probe above the cavity modifies the observed blue-shift value. Therefore, in order to quantify this resonance shift, we next recorded continuously the cavity transmittance spectra during a controlled probe scan above the whole structure. Preliminary, the probe was positioned at
272
B. Cluzel et al. / Photonics and Nanostructures – Fundamentals and Applications 9 (2011) 269–275
Fig. 4. (a) Cross-section of the electrical field intensity within the cavity at the resonance wavelength. (b) Cross-section of the standard perturbation of the cavity by the probe (according to the perturbation theory). (c, d) Different cross-sections of the extraordinary perturbation induced by the probe motion. The topographical lines corresponding to the probe motions associated to the optical measurements of (b–d) are plotted on the bottom.
1.5 mm away from the cavity centre and its motion was adjusted to be parallel to the waveguide axis. Then, as depicted in Fig. 6a, from t = 0 s, the probe is progressively approached from the cavity at a 0.2 Hz longitudinal scanning rate (around 1 mm/s). At t = t1, i.e. as soon as the probe scans the cavity centre, we set the probe motion to a single line scan along the cavity axis. This probe motion is maintained during a given time and at t = t2 the probe is instantaneously removed into the far-field. During all this procedure, the transmittance spectra were recorded continuously at a rate of 0.3 Hz. The transmittance spectrum evolution as a function of the probe scan is presented on a movie available
online and is summarized in Fig. 6c while the cavity resonance wavelength evolution extracted from the movie is plotted in Fig. 6b. All the recorded spectra exhibit a single resonance with high frequency oscillations due to Fabry Perot interferences between the ridge waveguide cleaved facets. As long as the probe is scanned out of the optical near-field of the nanocavity, the resonance stays unchanged. As soon as the probe penetrates the optical near-field of the cavity, the resonance wavelength starts blue-shifting. For t < t1, the cavity resonance decreases from 1585.4 nm to 1576 nm. Next, for t > t1, when the probe scan is blocked along a single axis centred on the cavity, the resonance wavelength continuously decreases to reach a
B. Cluzel et al. / Photonics and Nanostructures – Fundamentals and Applications 9 (2011) 269–275
273
Fig. 5. (a) Schematic view of the experiment: the near-field probe is maintained in the optical near-field of the cavity for a given waiting time tw. At t = 0 s, the probe is removed instantaneously into the far-field while recording for t > 0 s, the cavity transmittance temporal evolution. (b) Experimental measurements recorded for different waiting times: tw = 10, 30, 60 and 120 s.
final equilibrium value of 1569.4 nm at t1 + 300 s. For longer time, as long as the probe scan is maintained above the cavity, this resonance wavelength stays constant. Finally, we removed instantaneously the probe from the cavity optical near-field at t = t2 and clearly observed that the resonance wavelength increased progressively to reach its initial value at t2 + 150 s. At last, since the cavity peak transmittance as well as its resonance wavelength, come back to their initial values at the end of the experiment, the cavity was not damaged during the reported experiments.
3.5. Summary of the observations First, we found that the probe cavity interaction results in an extraordinary blue-shift of the cavity resonance wavelength, instead of a red-shift as expected from perturbation theory [13,14]. Next, the maximal blue-shift value is measured to be dl/l 1% (dl = 15 nm), an ‘‘extraordinary’’ value to be compared to the red-shift values of dl/l 0.01% (dl 0.2 nm) usually achieved with a similar probe. In addition, we find that a dynamical equilibrium state, which depends on the probe motion above the cavity, can be obtained. At last, the equilibrium state needs several tens of seconds to be achieved, which are
extraordinary long for an interaction relying only on a pure electromagnetic phenomenon. 4. Possible explanations In order to explain these unexpected results, one may consider several hypotheses. Assuming that the cavity length is not squeezed during the experiment the resonance shift is either due to a relative change of the refractive index dn/n 1% or to a change in the filling of the holes. Such refractive index change can be associated to either an electro-optic or a thermo-optic effect, which are the two dominant effects in silicon. We believe that the electro-optic effect is not involved for two strong reasons: - the time constant for the plasma dispersion effect in silicon is typically a few ns. - The density of carriers needed for such index change is in the excess of 1019 cm 3. Therefore, with such long time constants, thermooptical effects are more likely to be involved. However, in light of the temperature dependence of silicon refractive index [18–20], the relative temperature change corresponding to the reported dn/n 1% is higher than 50%, which corresponds to a cavity cooling
274
B. Cluzel et al. / Photonics and Nanostructures – Fundamentals and Applications 9 (2011) 269–275
Fig. 6. (a) Schematic view of the experiment. The cavity transmittance spectra are recorded during a probe scan: for t = 0 to t1, the probe is approached from the cavity, for t1 < t < t2, the probe motion is blocked to a single axis scan centred on the cavity. At t = t2, the probe is removed into the far field. (b) Temporal evolution of the cavity resonance wavelength extracted from the transmittance spectra recorded during the probe scan. The corresponding movie is available online. (c) Typical transmittance spectrum snapshots at different time during the experiment.
from an initial temperature of 300 K to a temperature lower than 150 K. Another possible explanation compatible with the wavelength shift amplitude may be the presence of trapped water inside the holes of the mirrors of the nanocavity which are affected by the motion of the near field probe on the top of the cavity. Even if we cannot be conclusive on this explanation it seems the most attractive. Firstly, the silica near-field probes are well-known to be hydrophilic and scanning the probe above holes filled with water could result in water being disturbed by the probe. Next, an FDTD simulation of a cavity, assuming that all the holes of the mirror are filled with water (n = 1.3) shows a 19-nm red shift of the resonance wavelength, in comparison to the same cavity without water. Then, measuring experimentally the transmit-
tance of an immersed nanocavity which does not exhibit the extraordinary shift leads to the observation of a 13 nm red-shift of its resonance wavelength. Those 3 values are all of the same order but the discrepancies are difficult to explain. Other works [21–23] have also shown that infiltration of liquids in holes can be used to tune photonic structures. However, these infiltration experiments have a time response of less than a second which lacks to explain our response time which can be up to 2 min. In addition, the experiment where the tip stays motionless on the centre of the cavity (Fig. 5) is hardly explained with this nanofluidics approach even if the few nanometre lateral oscillations of the probe due to the shear-force back could lead to weak amplitude water displacement at the surface of the nanocavity. Finally, the fact that the ‘‘extraordinary’’ regime is suppressed by a diluted HF etching (10 s in 1% HF
B. Cluzel et al. / Photonics and Nanostructures – Fundamentals and Applications 9 (2011) 269–275
solution) of the natural oxidization layer of silicon surface implies that this surface layer is deeply related to the reported phenomenon. Thus, combined to the fact that that the cavities are produced by dry etching without any wet etching post processing like HF, the presence of surface bonds which can trap water could explained the presence of water on the surface. Again another interpretation could rely on the presence of trapped electrostatic surface charges which could be created by the friction of the near-field probe on the surface. The presence of such charges could lead the existence of a strong electrostatic field able to strain the cavity by electrostriction. In this case, the dynamics of trapping and release of such charges [24] could agree with the time scale of the reported phenomenon. Against this last possibility remains the fact that the optical mode needs to be present for the effect to take place. Independently of the real causes, such an experiment shows that the preparation of a sample’s surface could have a tremendous influence in the field of NOEMS and nanophotonics. 5. Conclusions In conclusion, we reported a near-field interaction between a nanometric probe and a small-volume nanocavity, which allows the dynamical tuning of the cavity resonance over an extraordinarily broad spectral range. In light of all the experimental results, we were not able to identify rigorously the fundamental reasons of the observed interaction. Pure electro-thermo-optics effects as well as nanofluidic interactions with the nearfield probe do not seem to fully and satisfactorily explain all the experimental results. The most important conclusion of this experiment is that the surface preparation can have a drastic effect on NOEMS and consequently, we strongly believe that a cross-disciplinary analysis is required for further investigations of the reported phenomenon. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/ j.photonics.2011.05.002.
275
References [1] F. de Fornel, Evanescent Waves from Newtonian to Atomic Optics, Springer Verlag, Heidelberg, 2001. [2] L. Novotny, B. Hecht, Principles of Nano-Optics, Cambridge University Press, 2006. [3] P.T. Rakich, M.S. Dahlem, S. Tandon, M. Ibanescu, M. Soljacic, G.S. Petrtich, J.D. Joannopulos, L.A. Kolodziejski, E.P. Ippen, Nat. Mater. 5 (2006) 93. [4] N. Louvion, D. Ge´rard, J. Mouette, F. de Fornel, C. Seassal, X. Letartre, A. Rahmani, S. Callard, Phys. Rev. Lett. 94 (2005) 113907. [5] M.L.M. Balistreri, H. Gersen, J.P. Korterik, L. Kuipers, N.F. van Hulst, Science 294 (2001) 1080. [6] D.M. Eigler, E.K. Schweizer, Nature 344 (1990) 524. [7] M.F. Crommie, C.P. Lutz, D.M. Eigler, Science 262 (1993) 218. [8] R.D. Piner, et al. Science 283 (1999) 661. [9] T.R. Ramachandran, C. Baur, A. Bugacov, A. Madhukar, B.E. Koel, A. Requicha, G. Gazen, Nanotechnology 9 (1998) 237. [10] L. Lalouat, B. Cluzel, P. Velha, E. Picard, D. Peyrade, J.P. Hugonin, P. Lalanne, E. Hadji, F. de Fornel, Phys. Rev. B 76 (2007) 041102. [11] A.F. Koenderink, M. Kafesaki, B.C. Buchler, V. Sandoghdar, Phys. Rev. Lett. 95 (2005) 153904. [12] S. Mujumdar, A.F. Koenderink, T. Su¨nner, B.C. Buchler, M. Kamp, A. Forchel, V. Sandoghdar, Opt. Express 15 (2007) 17214. [13] B. Cluzel, L. Lalouat, P. Velha, E. Picard, D. Peyrade, J.C. Rodier, T. Charvolin, P. Lalanne, F. de Fornel, E. Hadji, Opt. Express 16 (2008) 279. [14] L. Lalouat, B. Cluzel, F. de Fornel, P. Velha, P. Lalanne, D. Peyrade, E. Picard, T. Charvolin, E. Hadji, App. Phys. Lett. 92 (2008) 111111. [15] C. Sauvan, G. Lecamp, P. Lalanne, J.P. Hugonin, Opt. Express 13 (2005) 245. [16] P. Velha, E. Picard, T. Charvolin, E. Hadji, J.C. Rodier, P. Lalanne, D. Peyrade, Opt. Express 15 (2007) 16090. [17] E. Silberstein, P. Lalanne, J.P. Hugonin, Q. Cao, J. Opt. Soc. Am. A 18 (2001) 2865. [18] J.A. McCaulley, V.M. Donnelly, M. Vernon, I. Taha, Phys. Rev. B 49 (1994) 7408. [19] B.J. Frey, D.B. Leviton, T.J. Madison, Proc. SPIE 7273 (2006). [20] M.W. Pruessner, T.H. Stievater, M.S. Ferraro, W.S. Rabinovich, Opt. Express 15 (2007) 7757. [21] S. Vignolini, F. Riboli, F. Intonti, M. Belotti, M. Gurioli, Y. Chen, M. Colocci, L.C. Andreani, W.S. Diederik, Phys. Rev. E 78 (2008) 45603. [22] M. Ebnali-Heidari, C. Grillet, C. Monat, B.J. Eggleton, Opt. Express 17 (2009) 1628. [23] C.L. Smith, U. Bog, S. Tomljenovic, M.W. Lee, D.K. Wu, L. O’Faolain, C. Monat, C. Grillet, T.F. Krauss, C. Karnutsch, R.C. McPhedran, B.J. Eggleton, Opt. Express 16 (2008) 15887. [24] J.S. Milne, A.C.H. Rowe, S. Arscott, Ch. Renner, Phys. Rev. Lett. 105 (22) (2010) 226802.