Electrically Tunable Organic Distributed Feedback Lasers Embedding Nonlinear Optical Molecules


  • Andrea Camposeo,

    Corresponding author
    1. National Nanotechnology Laboratory of Istituto Nanoscienze-CNR, Università del Salento, via Arnesano, I-73100 Lecce, Italy
    2. Center for Biomolecular Nanotechnologies@UNILE, Istituto Italiano di Tecnologia, via Barsanti, I-73010 Arnesano, Italy
    • National Nanotechnology Laboratory of Istituto Nanoscienze-CNR, Università del Salento, via Arnesano, I-73100 Lecce, Italy.
    Search for more papers by this author
  • Pompilio Del Carro,

    1. National Nanotechnology Laboratory of Istituto Nanoscienze-CNR, Università del Salento, via Arnesano, I-73100 Lecce, Italy
    Search for more papers by this author
  • Luana Persano,

    1. National Nanotechnology Laboratory of Istituto Nanoscienze-CNR, Università del Salento, via Arnesano, I-73100 Lecce, Italy
    2. Center for Biomolecular Nanotechnologies@UNILE, Istituto Italiano di Tecnologia, via Barsanti, I-73010 Arnesano, Italy
    Search for more papers by this author
  • Dario Pisignano

    Corresponding author
    1. National Nanotechnology Laboratory of Istituto Nanoscienze-CNR, Università del Salento, via Arnesano, I-73100 Lecce, Italy
    2. Center for Biomolecular Nanotechnologies@UNILE, Istituto Italiano di Tecnologia, via Barsanti, I-73010 Arnesano, Italy
    3. Dipartimento di Matematica e Fisica “Ennio De Giorgi”, Università del Salento, via Arnesano, I-73100 Lecce, Italy
    • National Nanotechnology Laboratory of Istituto Nanoscienze-CNR, Università del Salento, via Arnesano, I-73100 Lecce, Italy.
    Search for more papers by this author


original image

The emission of polymer distributed feedback lasers is tuned by integrating nonlinear optical organics. The nonlinearity allows reversible tuning of the amplified spontaneous emission of a conjugated polymer by 30 nm, and the lasing wavelength by 17 nm, with a lasing tunability coefficient of 0.17 nm/V.

Organic light-emitting compounds have revolutionized photonics and optoelectronics, opening the way to devices and applications hardly accessible by using inorganic materials. Color tunability, mechanical flexibility, large stimulated-emission cross sections, low costs and solution-based processing make active organic materials well-suited to complement inorganic photonic technologies. Using organic semiconductors in laser devices represents the most fascinating and debated field.1–3 The interest towards organic lasers relies on their potential application to optical communications,3 optical amplification,4 chemical sensing,5 and gain switching.6 However, the use of organic active media in laser devices is still delayed by the need of external sources for optical pumping and by the short operational lifetime. Recent steps forward include improved lifetime by proper encapsulation,7 hybrid devices exploiting inorganic excitation sources (diode lasers or light emitting diodes, LEDs)8, 9 and lasing from conjugated polymer films by two-photon pumping.10 At the same time, the broad gain bandwidths of organics would render laser devices ideal for tunable spectroscopy, particularly in the ultraviolet (UV) and visible spectral range where available laser sources mostly work at discrete wavelengths and with limited tunability. In this framework, organic lasers, for instance used to excite fluorescent biomarkers,11 can be an appealing and cheap alternative to conventional inorganic and gas laser, especially for applications where disposable, compact sources are required.12 Electrical tunability can represent a breakthrough for applying hybrid lasers as excitation sources in microfluidic lab-on-chips. In these systems, the integration of electrically tunable lasers would greatly enhance both detection sensitivity, by modulation of the laser intensity or frequency, and capability of parallel optical processing by multi-wavelength excitation.13

The emission of organic distributed feedback (DFB) lasers has been previously tuned by mechanical stretching,14 by exploiting a “wedge shape” active15 or intermediate high index layer,16 or by photoisomerizable azo-polymers.17 The basic idea of all these approaches is the variation of the DFB period (Λ) or of other geometric characteristics and, consequently, of the effective refractive index (neff), since the emission wavelength (λ) is given by the Bragg condition, mλ = 2neffΛ, where m is the diffraction order. Electrical tunability is reported in liquid crystal lasers,18–20 by using a liquid crystal as cladding layer,21 or an electroactive substrate.22 Most of these methods allow continuous and reversible tuning of the emission wavelength within a range <10 nm, whereas a remarkable tunability over 47 nm has been achieved in Ref. 22. Some approaches require moveable parts with potential problems of integration and handling, together with a smart management of the surface interlayers in order to prevent loss of adhesion.

In this Communication, we introduce a method to tune the emission, based on the integration of a nonlinear optical (NLO)-organic layer into a solid state, planar, polymer laser. The refractive index of NLO chromophores can be controlled by an external electric field, Eext. Second order effects, observed in optical crystals, display a linear dependence on Eext (linear electro-optic effect), whereas third-order nonlinear phenomena take place even in isotropic systems and are characterized by a quadratic dependence of the refractive index on Eext (quadratic electro-optic effect).23 Moreover, the use of NLO molecules allows to control the electromagnetic field at high speed (GHz frequencies), since no mechanical moving parts are involved.24

Our DFB device, schematized in Figure 1, comprises a layer of the prototype conjugated polymer poly[2-methoxy-5-(2-ethylhexyloxy)-1,4-phenylenevinylene] (MEH-PPV) as active material,25, 26 and a thin, optically transparent, layer of 9-ethyl-3-carbazolecarboxaldehyde-N-methyl-N-phenylhydrazone (NLO chromophore, details in the Supporting Information). In this system, the electrically induced change of the refractive index is expected to be mainly driven by the quadratic, off-resonant, DC electro-optic Kerr effect, since the spin-cast NLO layer is largely isotropic. Although third-order nonlinear effects are very weak in many materials, this approach is very advantageous for tuning organic lasers, since it does not need poling of the NLO layer (frequently used in linear Kerr materials), that usually requires high temperatures detrimental for the gain film.27 In our devices, the DC electric field is applied through Au interdigitated finger electrodes, providing a local electric field of the order of 106 V m−1. This geometry is fully compatible with light-emitting field effect transistors (FETs), that represent one of the most promising platforms for achieving lasing in organic semiconductors by electrical injection.3, 28, 29 The refractive index of the NLO layer under the applied electric field can be approximated as:23n = nNLO + 3χ(3)E2ext, where nNLO is the refractive index without external field (nNLO = 1.797 at λ = 600 nm, see Supp. Information) and χ(3) is the third-order dielectric susceptibility.

Figure 1.

a,b) SEM images of the tunable DFB device cross-section along the x-z (a) and x-y (b) planes, respectively. Scalebars: 500 nm (a) and 3 μm (b). c) Schematics of the DFB laser device structure comprising a layer of NLO chromophores. The DC electric field is applied through interdigitated finger Cr/Au electrodes (thickness 10 nm/70 nm, respectively), deposited on a Si/SiO2 substrate.

Firstly, we characterize the emission from the unstructured slab waveguide composed by the two organic layers deposited on Si/SiO2 substrates and integrating the metal electrodes. Upon increasing the fluence of the pumping UV laser beam, the emission collapses in a sharp peak (width of 11 nm), as typical of amplified spontaneous emission (ASE, Figure 2a). The peak wavelength (λASE,V = 0 = 617 nm) without applied voltage is determined by the slab waveguide mode structure. In particular, we chose the thickness of the active layer such to determine a cutoff wavelength, for light guided in the slab, within the gain bandwidth of MEH-PPV.26 This approach allows to tune effectively the emission near the cutoff wavelength.30 To tune ASE, we use the change of refractive index of the NLO layer, and consequently of the cutoff wavelength of the related waveguide, by the external electric field. Indeed, upon increasing the applied voltage we observe a blue-shift of ASE (25 nm at 43 V). In addition, this shift is fully reversible, and the ASE peak wavelength returns to its initial value by switching off the applied voltage. Figure 2b displays the dependence of the ASE shift on the applied voltage, which reproduces well the trend of the cutoff wavelength (inset of Figure 2b). The calculation of the cutoff wavelength is performed by modeling the devices as a four-layer slab waveguide, composed by the SiO2 substrate, the NLO layer, the MEH-PPV active layer, and air. A cutoff condition is achieved when the effective refractive index of the guided mode equals the NLO refractive index (neff = nNLO). Upon approaching this condition, the propagating mode is no more confined in the gain layer, and consequently ASE and lasing are less favored. By using this condition one can relate the cutoff wavelength to the external electric field (Eext), taking into account the electric field dependence of the NLO refractive index (see Supp. Information, Equations S1 and S2). For comparison, ASE from a bare MEH-PPV film, also displayed in Figure 2b, shows a limited blue-shift (6 nm), attributable to nonlinear properties of the conjugated polymer layer.31, 32 These results demonstrate that NLO layers allow to tune the emission wavelength in slab waveguides made by gain polymers.

Figure 2.

a) A sequence of ASE spectra collected at different applied voltages. By applying a voltage of 43 V, the ASE peak blue-shifts by about 25 nm (middle spectrum, red line), whereas it returns to the its starting wavelength (bottom, black curve) if the voltage is lowered again to a zero value (top, black curve). b) Measured shift of the ASE peak wavelength vs applied voltage, in a slab waveguide with (red circles) and without the NLO layer (blue squares). The dashed line is a fit to the data by a quadratic power law function, as expected due the electric field dependence of nNLO23 and, consequently, of the cutoff wavelength (See Supp. Information for details) whereas the continuous curve is a guide for the eyes. Inset: calculated cutoff wavelength vs applied electric field, E. For the calculation χ(3) of the order of 10−16 m2 V−2 is assumed.23

We apply this approach to tune the emission of organic DFB lasers. Nanopatterns to define the DFB cavities are produced by room-temperature nanoimprint lithography (RT-NIL),25, 33 using a rigid master template to deform the surface of the polymeric film by applying a pressure. The patterning process is entirely performed at room temperature in air, without surface physico-chemical treatments or exposures to light or particle beams as in conventional lithographic methods, thus fully preserving the emission and gain of the polymeric layer.25 The surface topography of an imprinted organic bilayer, measured by atomic force microscopy (AFM) is shown in Figure 3. The bilayer is composed by molecular systems with different molecular weights and viscoelastic properties, i.e. MEH-PPV on the top and the low-molar-mass NLO chromophore underneath. The master pattern is faithfully transferred in a single step, notwithstanding the height corrugation introduced by the presence of the interdigitated electrodes (Figure 3b). This indicates how robust the patterning method is against disuniformities of the thickness or rheology of imprinted samples. In RT-NIL of organic bilayers with constituents characterized by different rheology behavior and free volume fractions, the deformation dynamics can be different from that of conventional, single-layer imprinting. In fact, due to the molecular weight and steric hindrance dependence of the shear-compliance temporal plateau, after which terminal-flow hence irreversible patterning is accomplished, the low-molar mass layer is expected to be deformed more easily than the polymer.34 In our bilayer system, this effect can result in a conformal sinking and free volume contraction of the bottom NLO film, in turn assisting the pattern transfer onto the top polymer film. Further experiments to clarify the physical mechanisms affecting the complex rheology of the bilayer are currently underway in our laboratory.

Figure 3.

a,b) Morphology of the substrate with interdigitated electrodes deposited on a Si/SiO2 substrate, before the deposition of the organic layers (a) and of the active organic bilayer (NLO/MEH-PPV) nanostructured by RT-NIL (b), as measured by AFM. A gap of about 5 μm is present between adjacent interdigitated electrodes. The DFB grating period is Λ = 400 nm. c–e) Planar AFM view of the imprinted bilayer and height profiles along perpendicular directions.

In DFB lasers, we firstly consider the eventual effect of the metal electrodes. Figures 4a,b display the lasing emission (Light-out) versus the pumping fluence (Light-in) (L-L plot) for devices without and with underlying electrodes, respectively. The presence of the metal electrodes slightly increases the pumping threshold (from 40 to 60 μJ cm−2) and the emission linewidth by 10%, and decreases the polarization ratio of the emitted light by 30%. These variations are expected due to absorption losses induced by the metal electrodes and to the changes of the overall waveguiding properties of the DFB structure,28, 35 although spatial overlap of optical modes with electrodes is reduced by the presence of the NLO layer. We find that the inclusion of the NLO layer does not significantly alter the device threshold and lasing linewidth.25 The emission wavelength of the electrically-contacted devices (photograph in Figure 4c) embedding the NLO layer can be shifted continuously, during operation, up to a maximum blue-shift of 17 nm obtained applying a voltage of 100 V (Figure 4d). This result corresponds to a tunability coefficient of 0.17 nm/V, almost twice the best performances reported by other electrically-driven tunability concepts.18–22 Lasing occurs in the low-energy tail of the gain peak, evidencing that the shift of the waveguide cutoff wavelength is responsible of the observed tunability. Upon tuning, we find a slight increase of the lasing mode linewidth (inset of Figure 4d), since the used grating is less effective in providing feedback when the emission peak blue-shifts. Gratings with shorter period or two-dimensional cavity structures could likely provide a wider tunability range in order to exploit the full gain bandwidth of MEH-PPV.26

Figure 4.

a,b) L-L plot of organic DFB laser devices fabricated on Si/SiO2 substrates without (a) and with Cr/Au interdigitated electrodes (b). In the insets the polarized lasing spectra are displayed, evidencing a TE polarized lasing mode. The devices with (without) the metal electrodes display a pumping fluence of about 60 (40) μJ cm−2 and a polarization ratio (i.e. the ratio between the intensity of the TE and TM polarized lasing modes) of 12 (18), respectively. c) Photograph of the DFB device embedding the NLO layer and the interdigitated electrodes, connected to an external voltage controller. d) Emission wavelength tunability vs applied voltage. Inset: spectra collected at different voltages. The lasing linewidth increases from 2 nm at V = 0 to 2.6 nm at V = 100 V. The broad emission background and linewidth are mainly due to the quality of the imprinted structures and to the presence of competing Bragg modes typical of index-modulated DFB cavities.

In conclusion, we demonstrate continuous tuning of the wavelength of ASE of an organic slab waveguide and of lasing of a polymer DFB device, by embedding NLO molecules. These results demonstrate the possibility to change the emission wavelength of organic semiconductor lasers by an electric field, simply turning a knob, thus opening the route to the use of organic and hybrid organic-inorganic laser sources for analytical spectroscopy, on-chip biomedical diagnostics, high-sensitive chemical sensing etc. In perspective, this concept can be extended to other photonic and optoelectronic systems such as vertical cavity surface- emitting lasers, waveguides, organic LEDs and light-emitting FETs, where the integration of NLO layers and suitable electrode geometries can allow to tune, modulate and switch the emission intensity and wavelength by tailoring the effective refractive index.

Experimental Section

Device Fabrication: The DFB devices are fabricated on n-type Si substrates with a 800 nm thick layer of SiO2, chosen to favor optical confinement of guided modes into the gain layer. On top of a clean Si/SiO2 substrate, we firstly define the interdigitated Au electrodes (1000 fingers, on an overall area of 8 × 8 mm2) by photolithography and lift-off. To this aim, a 1.1 μm-thick film of positive photoresist (AZ5214E, Clariant) is spin cast onto the SiO2 surface and soft baked (100 °C). After UV exposure through a mask aligner (EVG 620) and resist development, a 10 nm adhesion layer of Cr and a 70 nm layer of Au are deposited onto the patterned resist by electron-beam evaporation. The electrodes are then created by soaking the wafer in acetone to lift-off the metal not adhering to the substrate.

The NLO/MEH-PPV bilayer is deposited directly by sequential spin-coating. The NLO chromophore (M.W. = 327) and MEH-PPV (M.W. = 1.8 × 103) are dissolved in chloroform (5 × 10−2 M) and toluene (4 × 10−6 M), respectively, and then deposited to form films of thickness of 130 nm and about 200 nm, respectively. The thickness of the NLO layer is chosen close to the height of electrodes, thus minimizing possible effects due to inhomogeneous electric field distribution as produced by the interdigitated metal fingers.

The DFB cavity is defined on the free surface of the active organic layer by RT-NIL, exploiting as mould a Si master fabricated by electron-beam lithography, lift-off and reactive-ion etching.25, 33 Transfer of the features of the Si master template (Λ = 400 nm, feature depth = 300 nm) onto the organic bilayer is performed by applying a pressure (∼100 MPa through a precision manual press) without the use of any anti-sticking layer. The RT-NIL process is performed in air at room temperature, thus preventing any degradation of the active layers.25

Device Characterization: Scanning electron microscopy is performed by using a Nova NanoSEM 450 system (FEI). The morphology of the imprinted organic bilayer is characterized by AFM, using a Multimode head equipped with a Nanoscope IIIa controller (Veeco), operating in tapping-mode with Si cantilevers (resonance frequency of about 275 kHz). ASE and lasing are obtained by pumping the devices by the third harmonic (λ = 355 nm) of a 3 ns Q-switched Nd:YAG laser (Spectra-Physics, repetition rate of 10 Hz). The excitation beam is tightly focused on the sample by a cylindrical lens into a rectangular excitation stripe (area about 10 × 0.5 mm2). The emission is collected by an optical fiber and spectrally dispersed by a monochromator (iHR320, Jobin Yvon) equipped with a charge-coupled device (Simphony, Jobin Yvon). All the optical measurements are carried at room temperature and in vacuum. For wavelength tunability, the voltage is applied to the interdigitated finger electrodes by a low-noise DC voltage source/meter instrument (Mod. 2400, Keithley Instruments).

Supporting Information

Supporting Information is available from the Wiley Online Library or from the author. Molecular structure and optical characterization of the used materials and details of waveguide modeling are included.


S. Tavazzi and G. Potente are gratefully acknowledged for the ellipsometric characterization of the NLO chromophore and for SEM, respectively. This work was partially supported by the Italian Minister of University and Research (MIUR) through the FIRB RBFR08DJZI “Futuro in Ricerca” project.