Single-step fabrication of surface waveguides in fused silica with few-cycle laser pulses
Federico J. Furch, W. Dieter Engel, Tobias Witting, Tobias Witting, Marc J. J. Vrakking, Alexandre Mermillod-Blondin
aa r X i v : . [ phy s i c s . a pp - ph ] A ug Single-step fabrication of surface waveguides infused silica with few-cycle laser pulses
Federico J. Furch , W. Dieter Engel , Tobias Witting , ArmandoPerez-Leija , Marc J. J. Vrakking , and AlexandreMermillod-Blondin Max-Born-Institut f¨ur Nichtlineare Optik undKurzzeitspektroskopie, Max-Born-Straße, D-12489 Berlin,Germany * Corresponding author: [email protected] 30, 2019
Abstract
Direct laser writing of surface waveguides with ultrashort pulses isa crucial achievement towards all-laser manufacturing of photonic inte-grated circuits sensitive to their environment. In this Letter, few-cyclelaser pulses (with a sub-10 fs duration) are used to produce subsurfacewaveguides in a non-doped, non-coated fused silica substrate. The fab-rication technique relies on laser-induced microdensification below thethreshold for nanopore formation. The optical losses of the fabricatedwaveguides are governed by the optical properties of the superstrate. Wehave measured losses ranging from less than 0.1 dB/mm (air superstrate)up to 2.8 dB/mm when immersion oil is applied on top of the waveguide.
Among the several fabrication techniques that may be used to producewaveguides in glass substrates [12], fs-laser direct writing [8] is especially at-tractive because it does not require a cleanroom environment, is cost-effective,and offers a high throughput [12]. Fs-laser writing enables full 3d waveguidefabrication in crystals [7], polymers [2] as well as glasses [10]. Although re-search efforts have been mostly concentrated on volume microprocessing, thedirect fabrication of surface waveguides would greatly extend the domain ofapplications of laser-written photonic integrated chips. The propagation of anoptical field in a surface waveguide implies the presence of an evanescent wavewhich can be exploited for refractive index sensing [19, 20], plasmonic excitationsensing [29] or Fourier-transform spectrometry [22]. However, using fs-laser di-rect writing on or near the surface of the host material is challenging. Attemptsin non-optimized glasses have resulted in surface swelling [6], cracking [19] and1blation [33]. In pure fused silica, the direct fabrication of near-surface waveg-uides with a good refractive index contrast has even been considered impossible[5]. The main reason is that fs-laser direct writing relies on pulse-to-pulse heataccumulation to induce controlled, localized heating of the substrate. After cool-ing, high-refractive index regions with light-guiding capabilities appear. In thisscheme, the magnitude of the stress load produced upon thermal relaxation is amajor limitation. Recently, several routes have been explored to enable surfacewaveguide photoinscription. One strategy consists in using toughened glasses asa host substrate [19, 20]. Another approach is to enhance the photosensitivity ofthe glass by doping with silver ions [1]. A third method relies on suppressing theair/dielectric interface by bonding a thin glass on top of the sample [5]. Becausethe bond is ensured by weak van der Waals attractive forces (only manifestingwhen the sample and the cover glass are in close contact), such a method mightonly be applicable to planar substrates. In this Letter, we describe the directfabrication of surface waveguides in fused silica with the help of few-cycle laserpulses. Our method does not rely on pulse-to-pulse heat accumulation, but in-stead is based on the type 1 laser-matter interaction regime [25] triggered ina grazing incidence irradiation scheme. Modifications induced in the type 1regime exhibit a smooth, uniform and positive refractive index change ∆ n , incontrast to type 2 modifications which are characterized by an important bire-fringence due to the presence of periodic nanogratings in the irradiated region.Furthermore, we demonstrate the possibility to control the propagation lossesby playing on the refractive index of the superstrate which opens a route for thelaser-assisted fabrication of non-hermitian microoptical systems [9].The experimental setup is depicted in Fig. 1. Few-cycle pulses (sub-10 fsduration, central wavelength at 800 nm) from a high repetition rate (400 kHz),non-collinear optical parametric amplifier [11] are focused on the surface of afused silica sample with a grazing incidence. The samples are parallelepipedicand polished to optical quality on all sides. In order to preserve the temporalstructure of the laser pulse in the focal plane, we use a gold-coated reflectiveobjective (numerical aperture 0.5) [26]. Such an irradiation scheme providestwo focii F and F formed by the upper and lower halves of the laser beam,respectively. The presence of a planar air/glass interface induces wavefrontdistortions and leads to the formation of an aberrated focus F [18] which isshifted and spread along the propagation axis. The corresponding laser-inducedmodifications are shown in Fig. 1(b). An analogous distortion happens in thetime domain with consequences on the temporal profile of the irradiation [31].Because the spatio-temporal distortions imparted on the lower half of the beamincrease with the amount of propagation in the sample, the irradiation char-acteristics vary when translating the sample along the propagation axis. Inorder to limit the irradiation to the part of the beam traveling through air only,the lower half of the laser beam was blocked. Furthermore, the entrance of thebeam diameter was slightly reduced with the help of an iris, providing an overalltransmission of 0.17 for the focusing unit. The fabrication of a single line-shapedmicrostructure is then straightforward. It suffices to irradiate the surface of thesubstrate continuously from edge to edge by translating the sample with the2igure 1: (a): Cross-sectional view of the experimental setup used for lasermicroprocessing at grazing incidence. The part of the laser beam representedin bright red propagates through air only and forms the focus F . The beamblocker can be put in the beam path to prevent the formation of the aberratedfocus F . (b): Phase-contrast microscopy (PCM) picture of the sample’s topsurface showing the refractive index distribution resulting from F and F .help of a stepper motor. In what follows, the laser pulse energy was constantwith a value of E = 530 nJ after the microscope objective and the speed of thestepper motor was 60 µ m s − . The laser-induced microstructures had a lengthof 10 mm (i.e. the width of the sample). For higher pulse energies, intense laserlight scattering occurred and the irradiated region exhibited a mix of negativeand positive refractive index changes (not shown). These features are indicativeof the type 2 interaction regime characterized by the formation of nanopores[4, 25].In Fig. 2(a) we show the laser footprint on the exit facet after sample pol-ishing, using optical transmission microscopy (numerical aperture 0.9). Thelaser-induced microstructure has a width w ≈ µ m and a height h ≈ µ m.Diagnostics of the top surface with an atomic force microscope (AFM) are pre-sented in Fig. 2(b), and reveal the presence of a shallow surface topographyvariation ( < −
10 nm, about 430 nm FWHM) on top of the laser-induced mi-crostructure. The negative sign of the topology variation is indicative of avolume reduction (and hence a density increase) in the irradiated area, and isthe opposite to what happens when microprocessing is performed in the type 2regime, with longer (35 fs) and more energetic (about 2 µ J) laser pulses wheresurface swelling as high as 250 nm was measured [6]. An inversion in the sign ofthe surface topography is consistent with recent observations reporting a volumereduction of glass cantilevers irradiated in the type 1 regime and a net volumeincrease of cantilevers irradiated in the type 2 regime [4]. We emphasize that theabsence of material re-deposition in the AFM pictures hints towards a purelynon-ablative process. The phase shift distribution ∆ φ across the laser-inducedmicrostructure presented in Fig. 2(c) was measured by spatial light interferencemicroscopy [35]. As expected from a local density increase, ∆ φ is positive in theirradiated volume [32]. The corresponding spatial average of the laser-induced3efractive index change ∆ n = λ c ∆ φ πh ≈ . λ c =550 nm is the cen-tral wavelength of the illumination source (an halogen light bulb in our case), h = 6 µ m is defined in Fig. 2(a) and ∆ φ = 0 .
43 rad is the phase shift measuredat the center of the microstructure. We emphasize that the magnitude of ∆ n obtained exceeds the value of 10 − − − usually measured in the bulk for type1 modifications [25]. D2 (cid:1) -1 ] I n t e n s i t y [ a r b . un i t s ] Substrate (fused silica)
Laser-induced microstructure (end facet)Air (cid:2) m(a)AFM 2 (cid:0) m Height [nm]
430 nm(b)
Phase shift [rad.]
QPCM 2 (cid:3) m(c)(d) h ~ 6 (cid:4) mw ~ 2 (cid:5) m Figure 2: Characterization of the laser-induced optical structures. (a): Sideview of the sample acquired with an optical transmission microscope. Thesample is illuminated with an halogen lamp. (b): Surface topography (top view)measured with an atomic force microscope (AFM).(c): Phase shift across themicrostructure, measured with a spatial light interference microscope (SLIM).(d): Micro-Raman investigations of the irradiated zone (purple line) and of thepristine sample (grey line). The spectra were normalized with respect to the ω . µ m. The spectra were normalized with respect to the amplitude of the ω − ) inthe irradiated volume, confirming a local laser-induced compaction [3, 4]. Thesemicro-Raman measurements were carried out on the top surface of the sampleand might not necessarily correspond to the maximum of compaction, presum-ably located in the center of the laser-induced microstructure (i.e. ≈ µ m awayfrom the surface).In order to examine the ability of the microstructures to guide light at op-tical frequencies, the fundamental mode of a CW He-Ne beam ( λ = 633 nm)was focused onto the entrance facet of the laser-induced microstructures with amicroscope objective (numerical aperture 0.42). A second microscope objective(Olympus MPlan, 100x, numerical aperture 0.9) used in combination with atube lens (focal length of 200 mm) formed a 111-fold magnified image of theend facet on a camera sensor. Figure 3(a) shows the obtained output intensitydistribution. It demonstrates that these laser-induced microstructures supportoptical waveguiding at 633 nm. A modal analysis in the directions parallel ( p − )and perpendicular ( s − ) to the surface of the sample (see Fig. 3(b)) provides amode field diameter (MFD, defined as the 1 /e decay of the maximum modeintensity), of 4.3 and 6.0 µ m in the p − and s − directions, respectively. Outof the center region, the guided mode intensity decays exponentially [16] withdecay constants of ≈ . µ m and 0 . µ m in glass and in air, respectively. Wechecked that the optical transfer function of the microscope objective did notsignificantly influence these values by applying a deconvolution algorithm to thecurves shown in Fig. 3(b). The point spread function used for the deconvolutionwas estimated numerically based on the model of Gibson and Lanni [13, 23].The influence of the input polarization was examined by placing the waveg-uide between a polarizer and an analyzer. The transmission of the waveguidewas measured as a function of the relative angle between the polarizer and theanalyzer [see Fig. 3(b)]. The polarization is maintained for input fields witha linear polarization in the s- and p- directions (see thick transparent linesin Fig. 3(b) left). However, an input field with a linear input polarization inanother direction becomes elliptically polarized upon propagation in the waveg-uide, which indicates that s- and p- polarized fields have different propagationconstants.Having confirmed the waveguiding capabilities of the laser-induced microstruc-tures and their polarization-maintaining properties, we now examine the pos-sibility to control the propagation losses by varying the refractive index of thesuperstrate n s . By applying the optimum end-fire coupling method [15], losses < . − were measured for n s = 1 .
00 (in air). The losses obtained whenapplying a drop of immersion oil on top of the waveguide are shown in Fig. 4.The diameter of the oil droplet was controlled by using a graduated microsy-ringe.A first order exponential fit of the experimental data indicates that the5igure 3: (a): Near-field intensity profile at the exit of the laser-induced mi-crostructure at a wavelength of 633 nm. (b): Mode-field analysis. The regionsin gray correspond to the size of the waveguide core deduced from Fig. 2(a).The dotted lines are first-order exponential fits of the experimental data. MFD:mode field diameter defined as the 1 /e decay of the maximum intensity. (c):Transmitted intensity as a function of the analyzer angle for linear input po-larizations in the p- and s- directions (left) and for an arbitrary linear inputpolarization (right). The dark thin lines represent the polarization of the beamafter propagation in the waveguide and the thick, semi transparent lines repre-sent the polarization of the input beam.6 il droplet diameter [mm] T r a n s m i tt e d i n t e n s i t y [ a r b . un i t s ] ii iii iv v WaveguideOil droplet
Figure 4: Evolution of the optical transmission at 633 nm as a droplet of im-mersion oil ( n oil = 1.516) with a variable diameter is placed on top of thelaser-induced surface waveguide. The dotted line represents a fit of the exper-imental data using a single exponential decay. The presence of oil induces anattenuation of 2.83 +0 . − . dB mm − .superstrate-induced leakage is as strong as 2 . +0 . − . dB mm − . The uncertain-ties were obtained from fits of the lower and upper bounds of the experimentalvalues. In the so-called ray-optic approach of guided mode theory [34], the lightpropagating through a waveguide is described as a sum of oblique rays experi-encing total internal reflection at the boundary of the waveguide. Changing n s changes the conditions for total internal reflection. When n s increases, the min-imum angle for total internal reflection decreases and the steepest rays escapethe waveguide [27, 20].The large attenuation coefficient deduced from the numerical fit indicatesthat the optical structures are sensitive to the refractive index of their envi-ronment and can thus be employed as refractive index sensors, with significantpotential for lab-on-chip applications. These waveguides may for instance con-stitute the sensitive part of optical biosensors microsystems that are used forlabel-free bio-sensing [29]. The optical structures presented in this Letter mayalso be used as efficient photon/plasmon couplers to interface photonic andplasmonic architectures [14], or as the backbone for the fabrication of compactstationary-wave integrated Fourier-transform spectrometers [22]. Furthermore,the ability to control the losses in integrated waveguide configurations opens thedoor to new perspectives to manufacture non-Hermitian non-resonant photonicsystems in connection with exceptional points singularities [24].In this Letter we have presented a method to inscribe waveguides directlyon the surface of a pure fused silica substrate. For the results presented here,few-cycle pulses from a high repetition rate non-collinear optical parametricamplifier have been utilized. The extent to which longer pulses can be utilizedwill be the subject of future investigations. We note that stretching the pulse bylinear dispersion is not a viable option. The highly structured ultra broadbandspectrum quickly leads to a multi-pulse structure in the time domain, which7ould potentially change the dynamics of the ionization process. We also notethat new trends in non-linear pulse compression have the potential to bringfew-cycle pulse capabilities to lasers typically used in laser material processing[21].The results presented in this Letter represent the first demonstration ofwaveguides that are directly photoinduced on the surface of fused silica withoutthe need for pre- (e.g. deposition of a photosentive material or applicaton ofa cover slip) or post-processing of the target substrate. The optical structuresproduced correspond to type 1 modifications, support waveguiding at opticalfrequencies, possesses polarization-maintaining properties and exhibit a core re-fractive index change of ∆ n ≈ +0 .
006 on average. AFM measurements andMicro-Raman investigations indicate that laser-induced microcompaction is atthe origin of the observed refractive index change. The waveguides are sensitiveto their environment, extending the capability of the direct laser write methodto the rapid prototyping of compact optical non-Hermitian microsystems tak-ing advantage of all the well-known benefits (small size and weight, low powerconsumption, improved reliability and vibration sensitivity [17]) of integratedoptical devices.
Funding Information
Deutsche Forschungsgemeinschaft, Grants ME4427/1-1 and ME4427/1-2.
Acknowledgments
The authors thank J. Tomm and S. Schwirzke-Schaaf for their assistance withthe micro-Raman measurements.
Supplemental DocumentsReferences [1] Alain Abou Khalil, Jean-Philippe B´erub´e, Sylvain Danto, Jean-CharlesDesmoulin, Thierry Cardinal, Yannick Petit, R´eal Vall´ee, and Lionel Can-ioni. Direct laser writing of a new type of waveguides in silver containingglasses.
Scientific Reports , 7(1):11124, September 2017.[2] Alexandra Baum, Patricia J. Scully, Maria Basanta, C. L. Paul Thomas,Peter R. Fielden, Nicholas J. Goddard, Walter Perrie, and Paul R. Chalker.Photochemistry of refractive index structures in poly(methyl methacrylate)by femtosecond laser irradiation.
Opt. Lett. , 32(2):190–192, Jan 2007.[3] Y. Bellouard, E. Barthel, A. A. Said, M. Dugan, and P. Bado. Scanningthermal microscopy and raman analysis of bulk fused silica exposed to low-8nergy femtosecond laser pulses.
Opt. Express , 16(24):19520–19534, Nov2008.[4] Yves Bellouard, Audrey Champion, Benjamin McMillen, SebabrataMukherjee, Robert R. Thomson, Charles P´epin, Philippe Gillet, andYa Cheng. Stress-state manipulation in fused silica via femtosecond laserirradiation.
Optica , 3(12):1285–1293, Dec 2016.[5] Jean-Philippe B´erub´e and R´eal Vall´ee. Femtosecond laser direct inscriptionof surface skimming waveguides in bulk glass.
Opt. Lett. , 41(13):3074–3077,Jul 2016.[6] V. R. Bhardwaj, P. B. Corkum, D. M. Rayner, C. Hnatovsky, E. Simova,and R. S. Taylor. Stress in femtosecond-laser-written waveguides in fusedsilica.
Opt. Lett. , 29(12):1312–1314, Jun 2004.[7] Feng Chen and J. R. V´azquez de Aldana. Optical waveguides in crystallinedielectric materials produced by femtosecond-laser micromachining.
Laser& Photonics Reviews , 8(2):251–275, 2014.[8] K. M. Davis, K. Miura, N. Sugimoto, and K. Hirao. Writing waveguides inglass with a femtosecond laser.
Opt. Lett. , 21(21):1729–1731, 1996.[9] Ramy El-Ganainy, Mercedeh Khajavikhan, Demetrios N. Christodoulides,and Sahin K. Ozdemir. The dawn of non-hermitian optics.
CommunicationsPhysics , 2(1):37, March 2019.[10] C. Florea and K. A. Winick. Fabrication and characterization of photonicdevices directly written in glass using femtosecond laser pulses.
Journal ofLightwave Technology , 21(1):246–253, Jan 2003.[11] Federico J. Furch, Achut Giree, Felipe Morales, Alexandria Anderson,Yicheng Wang, Claus Peter Schulz, and Marc J. J. Vrakking. Close totransform-limited, few-cycle 12 µ j pulses at 400 khz for applications inultrafast spectroscopy. Opt. Express , 24(17):19293–19310, Aug 2016.[12] Andrea Chiappini Giancarlo C. Righini. Glass optical waveguides: a reviewof fabrication techniques.
Optical Engineering , 53:53 – 53 – 15, 2014.[13] Sarah Frisken Gibson and Frederick Lanni. Diffraction by a circular aper-ture as a model for three-dimensional optical microscopy.
J. Opt. Soc. Am.A , 6(9):1357–1367, Sep 1989.[14] Xin Guo, Min Qiu, Jiming Bao, Benjamin J. Wiley, Qing Yang, XiningZhang, Yaoguang Ma, Huakang Yu, and Limin Tong. Direct coupling ofplasmonic and photonic nanowires for hybrid nanophotonic componentsand circuits.
Nano Letters , 9(12):4515–4519, 2009. PMID: 19995088.[15] M. Haruna, Y. Segawa, and H. Nishihara. Nondestructive and simplemethod of optical-waveguide loss measurement with optimisation of end-fire coupling.
Electronics Letters , 28(17):1612–1613, 1992.916] Jonathan Hu and Curtis R. Menyuk. Understanding leaky modes: slabwaveguide revisited.
Adv. Opt. Photon. , 1(1):58–106, Jan 2009.[17] Robert G. Hunsperger.
Integrated Optics Theory and Technology 6th Edi-tion . Springer, 2009.[18] N. Huot, R. Stoian, A. Mermillod-Blondin, C. Mauclair, and E. Audouard.Analysis of the effects of spherical aberration on ultrafast laser-inducedrefractive index variation in glass.
Opt. Express , 15(19):12395–12408, SEP17 2007.[19] Jerome Lapointe, Mathieu Gagn´e, Ming-Jun Li, and Raman Kashyap.Making smart phones smarter with photonics.
Opt. Express , 22(13):15473–15483, Jun 2014.[20] Jerome Lapointe, Francois Parent, Elton Soares de Lima Filho, S´ebastienLoranger, and Raman Kashyap. Toward the integration of optical sen-sors in smartphone screens using femtosecond laser writing.
Opt. Lett. ,40(23):5654–5657, Dec 2015.[21] L. Lavenu, M. Natile, F. Guichard, Y. Zaouter, X. Delen, M. Hanna,E. Mottay, and P. Georges. Nonlinear pulse compression based on a gas-filled multipass cell.
Opt. Lett. , 43(10):2252–2255, May 2018.[22] Etienne Le Coarer, Sylvain Blaize, Pierre Benech, Ilan Stefanon, AlainMorand, Gilles Lerondel, Gregory Leblond, Pierre Kern, Jean Marc Fedeli,and Pascal Royer. Wavelength-scale stationary-wave integrated fourier-transform spectrometry.
Nat Photon , 1(8):473–478, August 2007.[23] Jizhou Li, Feng Xue, and Thierry Blu. Fast and accurate three-dimensionalpoint spread function computation for fluorescence microscopy.
J. Opt. Soc.Am. A , 34(6):1029–1034, Jun 2017.[24] Mohammad-Ali Miri and Andrea Al`u. Exceptional points in optics andphotonics.
Science , 363(6422), 2019.[25] K. Mishchik, C. D’Amico, P. K. Velpula, C. Mauclair, A. Boukenter,Y. Ouerdane, and R. Stoian. Ultrafast laser induced electronic andstructural modifications in bulk fused silica.
Journal of Applied Physics ,114(13):133502, 2013.[26] B. Piglosiewicz, D. Sadiq, M. Mascheck, S. Schmidt, M. Silies, P. Vasa, andC. Lienau. Ultrasmall bullets of light—focusing few-cycle light pulses tothe diffraction limit.
Opt. Express , 19(15):14451–14463, Jul 2011.[27] F. Rehouma, D. Persegol, and A. Kevorkian. Optical waveguides for evanes-cent field sensing.
Applied Physics Letters , 65(12):1477–1479, 1994.1028] R. Saavedra, M. Le´on, P. Martin, D. Jim´enez-Rey, R. Vila, S. Girard,A. Boukenter, and Y. Ouerdane. Raman measurements in silica glassesirradiated with energetic ions.
AIP Conference Proceedings , 1624(1):118–124, 2014.[29] B Sep´ulveda, J S´anchez del R´ıo, M Moreno, F J Blanco, K Mayora,C Dom´ınguez, and L M Lechuga. Optical biosensor microsystems basedon the integration of highly sensitive machzehnder interferometer devices.
Journal of Optics A: Pure and Applied Optics , 8(7):S561, 2006.[30] N. S. Shcheblanov, M. E. Povarnitsyn, K. N. Mishchik, and A. Tanguy. Ra-man spectroscopy of femtosecond multipulse irradiation of vitreous silica:Experiment and simulation.
Phys. Rev. B , 97:054106, Feb 2018.[31] Bangshan Sun, Patrick S. Salter, and Martin J. Booth. Effects of sampledispersion on ultrafast laser focusing.
J. Opt. Soc. Am. B , 32(7):1272–1280,Jul 2015.[32] C.Z. Tan, J. Arndt, and H.S. Xie. Optical properties of densified silicaglasses.
Physica B: Condensed Matter , 252(12):28 – 33, 1998.[33] Gustavo A. Torchia, Pablo F. Meil´an, Airan Rodenas, Daniel Jaque, CruzMendez, and Luis Roso. Femtosecond laser written surface waveguidesfabricated in nd:yag ceramics.
Opt. Express , 15(20):13266–13271, Oct 2007.[34] R. Ulrich and R. J. Martin. Geometrical optics in thin film light guides.
Appl. Opt. , 10(9):2077–2085, Sep 1971.[35] Zhuo Wang, Larry Millet, Mustafa Mir, Huafeng Ding, Sakulsuk Unaruno-tai, John Rogers, Martha U. Gillette, and Gabriel Popescu. Spatial lightinterference microscopy (slim).