Presentation outline 1 2 • Why DFB lasers with (LC-RWG) surface gratings? • Particularities of LC-RWG surface gratings • Application & target characteristics: ESA requirements • Epilayer structures adjusted for LC-RWG gratings • Coupling coefficient & modal selectivity • Requirements on coupling coefficient and fabrication limitations • End-mirror effects • Device fabrication & fabricated device characteristics • Towards improved device characteristcs • narrower spectral linewidth Narrow linewidth 894 nm DFB lasers with laterally-coupled ridge-waveguide surface gratings fabricated using nanoimprint lithography M. Dumitrescu, Dumitrescu, J. Telkkä Telkkälä, J. Karinen, A. Laakso, J. Viheriä Viheriälä,, T. Leinonen, J. Lyytikä Lyytikäinen, L. Toikkanen, M. Pessa Optoelectronics Research Centre, Tampere University of Technology, Tampere, Finland Development of Extremely Narrow-Band Semiconductor Distributed Feedback Laser Technology The long path from idea to product: idea (ne w structure, ne w technology) + acquiring the facilities + proof of conce pt experime nts + project proposal + detailed analysis + design + expe rime nts + (many) ite rations CAS 2010 Tuesday, Oct. 12, O2 CAS 2010 Tuesday, Oct. 12, O2 3 Why DFB lasers with LCLC-RWG surface gratings? Differences in coupling coefficient dependence on grating structure 4 3 4 DFB laser with a buried grating: Surface gratings: single growth and processing sweep => better yield and lower cost limited interaction between the carriers & the grating => more controllable performances => more stable (GaAs) devices reduced coupling coefficient DFB laser with a LC-RWG grating: Akiba et al. IEEE J. Quantum Electron. 19, 1052 (1983) smaller Γ, higher Δn => high k is achievable The conventional buried-grating fabrication techniques require overgrowth, which leads to problems with yield, performance and device cost. Fabricated using UV-based nano-imprint lithography (UV-NIL), which enables pattern resolutions beyond the limitations set by the diffraction and scattering for the conventional techniques. basic dimensions: t, W, D, grating parameters: m, Λ, γ CAS 2010 longitudinal parameters: L, R1, R2, d1, d2 Tuesday, Oct. 12, O2 The effects of the grating order and filling factor on SMSR are not discussed in the literature (since buried-grating devices have been used). Λ1/Λ = filling factor k sin( πmγ ) κ = 0 ⋅ (n22 − n12 ) ⋅ Γg ⋅ πm 2 neff => with surface gratings high filling factor (i.e. small trench width) should be targeted Both transverse and longitudinal mode discrimination have to be taken into account CAS 2010 Tuesday, Oct. 12, O2 6 Epilayer structure Application & target characteristics Active region (894 nm) 5 6 The 1955 Cesium Atomic Clock at the National Physical Laboratory, UK.: Cesium clocks measure frequency with an accuracy of from 2 to 3 parts in 10 to the 14th, with an averaging time of 5 days. i.e. 0.00000000000002 Hz; this corresponds to a time measurement accuracy of one second in 1,400,000 years. It is the most accurate realization of a unit that mankind has yet achieved. For the emission at 894 nm GaInAs/GaAs QWs are too shallow and don’t have high enough gain for effective operation. In 1967, the 13th General Conference on Weights and Measures first defined the International System (SI) unit of time, the second, in terms of atomic time as the duration of 9,192,631,770 cycles of microwave light absorbed or emitted by the hyperfine transition of cesium-133 atoms in their ground state undisturbed by external fields. (a) Material gain simulated for 8 nm GaIn0.20 As/GaAs QW, and (b) the corresponding peak material gain compared with the variation reported in the literature (blue curve) The target: low-cost mass-produced atomic clock no larger than a sugar cube, which could run on an AA battery, yet accurate to one second in 300 years. The competition: National Institute of Standards and Technology (NIST), Boulder, USA, under the Defense Advanced Research Projects Agency’s (DARPA’s) Chip-Scale Atomic Clock (CSAC) program Cesium atomic clocks (ESA project for developing narrow linewidth DFB lasers) emission at 894.6 nm (tunable ± 2nm) (Cesium D1 line) => grating period Λ ≈ 140, 280, 420 nm output power ≥ 15 mW (@ Ibias < 100 mA) single mode CW operation with SMSR > 25 dB spectral linewidth << 1.5 MHz 171,173 Yb 87,88Sr 40Ca+ 27Al+ 199,201 Hg Cooling lasers (1st, 2nd , ...) Auxiliary lasers Lattice lasers Probe lasers (a) The peak material gain simulated for 5 nm GaInAs/Al0.30GaAs, GaInAs/GaIn0.485 P, and GaInAs/GaAs QWs, and (b) for 5 nm GaInAs/GaIn0.485 P QWs at 300, 325 and 350 K GaInAs/AlGaAs or GaInAsP/GaAsP QWs would be deep enough, but the growth temperatures for GaInAs and AlGaAs are quite different and it is not recommended to use AlGaAs next to GaInAs QWs for reliability reasons . The experimental studies indicate that the group V elements ratio in GaInAsP/GaAsP QWs is not easily controllable. => GaInAs/GaInP active regions (GaInP can be used as etch-stop) CAS 2010 Tuesday, Oct. 12, O2 CAS 2010 Tuesday, Oct. 12, O2 7 8 Epilayer structure Epilayer structure Waveguide and cladding layers (for high grating coupling coefficient) 7 Threshold current (mA) Al0.6GaAs cladding Al0.7GaAs cladding 350 50 325 45 300 40 275 250 35 225 Higher Al-content in the cladding layer • Wider far-field • Narrower near-field → cladding thickness can be reduced (fabrication of the grating is easier), but coupling coefficient is smaller due to smaller achievable Γg 200 30 175 150 50 SQW κ= 100 150 8 200 250 300 25 Waveguide thickness (nm) k0 sin(πmγ ) ⋅ ( n22 − n12 ) ⋅ Γg ⋅ 2neff πm 500 0 50 100 150 200 250 Surface gratings have low coupling coefficients (??) 300 0,016 400 300 0,014 200 0,012 100 0 0,010 -100 Drastic changes of the epilayer structure in order to extend the optical field penetration into the cladding affect other critical parameters and small changes do not bring significant coupling coefficient improvement. 0,008 -200 ”Barrier reduction” layers added to GaAs-Al 0.7GaAs interfaces in order to improve current-voltage characteristics. The QW confinement factor is usually maximized for low threshold and high bandwidth but this affects the coupling coefficient. single GaInAs quantum well (QW) 120 nm GaInP waveguide (WG),etch-stop layer. 100 nm Al0.3Ga0.7 As barrier-reduction layers 1000 nm thick Al0.7Ga0.3 As cladding layers, 200 nm GaAs contact layer on the p-side. QW confinement factor Al0.5GaAs cladding 375 FF-FWHM (degrees) High QW confinement factor • Small threshold current 55 NF FWHM extension beyond the WG [nm] Near-field FWHM [nm] 400 0 50 100 150 200 250 Etching depth should be used to ensure adequate coupling coefficient 300 GaInP waveguide thickness [nm] InGaAs InP WS L InGaAlAs GRINSCH QDashes GRINSCH InGaAlAs InP DW ⇒Etching through the active region increases dramatically the losses ⇒Etching should be stopped above the active region (is this enough?) L CAS 2010 Tuesday, Oct. 12, O2 CAS 2010 Tuesday, Oct. 12, O2 9 10 Coupling coefficient calculation Coupling coefficient calculation (effect of approximations) 9 2 ∫ ∫ ∆ε ( x, y, z ) ⋅ Ψ ( x, y ) dxdy k0 − ∞− ∞ ⋅ ∞ ∞ 2neff 2 ∫ ∫ Ψ ( x, y) dxdy − ∞− ∞ ∞ ∞ κ= k0 2neff [ ∞ ∞ ∫ ∫n 2 1 −∞ −∞ ∞ ∫ (x, y)Ψ 2 ( x, y)dxdy ∞ 2 ∫ Ψ ( x, y)dxdy − ∞− ∞ sin(πmγ ) k0 ⋅ ( n22 − n12 ) ⋅ Γg ⋅ 2neff πm κ= 2 2 Conventional approximation ≈ k0 ⋅ (n2 − n1 ) ⋅ Γg ⋅ The ’standard’ formula for calculating the coupling coefficient is over-estimating the LC-RWG grating coupling coefficient 300 ] sin(ππmmγ ) ∆ε m ( x, y , z ) = n2 ( x, y ) − n1 ( x, y ) ⋅ ∫ ∫ n22 ( x, y )Ψ 2 ( x, y)dxdy sin(πmγ ) − ∞− ∞ ⋅ ⋅ − ∞ ∞ πm 2 ∫ ∫ Ψ ( x, y )dxdy −∞ −∞ κ= K2139LD980 with W=1.5 µm and D=0.5 µm Coupling coefficient (1/cm) ∞ ∞ κ= 10 Dielectric perturbation Δεm(x,y,z) for m-th order rectangular-shaped grating: According to coupled-wave theory: sin( πmγ ) πm => ove r-estimatimation k0 sin(πmγ ) 2 2 ⋅ ( neff ,2 − neff ,1 ) ⋅ 2 neff πm Crosslight's PICS3D 'Standard' formula Convolution approach + adjusted formula Conventional approach + adjusted formula 250 200 κ= 150 κ0 2 neff ( ) ⋅ n22 − n12 ⋅ Γg ⋅ sin( πmγ ) sin( πmγ ) ≈ k 0 ⋅ ( n 2 − n1 ) ⋅ Γg ⋅ πm πm over-estimatimation since 100 50 0 0 50 100 150 200 250 300 2n eff > ( n2 + n1 ) The adjusted formula can produce underestimation if the optical field is calculated separately for each grating section k sin(πmγ ) κ = 0 ⋅ ( n eff , 2 2 − neff ,1 2 ) ⋅ πm 2neff Thickness of the un-etched cladding (nm) Third-order LC-RWG grating coupling coefficient values calculated with the standard formula (stars), with the adjusted formula and conventional method (triangles), with the adjusted formula and convolution method (circles) and obtained with PICS3D (squares) neff , i = ∫Ψ 2 ( x, y ) ⋅ n 2 (x , y ) ⋅dx ⋅ dy i 2 ∫ Ψ (x, y) ⋅dx (neff2_conventional-neff1_conventional) < (neff2_convolution - neff1_convolution) unde r-estimatimation possible CAS 2010 (for certain dimensional ranges) Tuesday, Oct. 12, O2 CAS 2010 Tuesday, Oct. 12, O2 11 Transverse modal selectivity SMSR calculation (coupling coefficient discrimination) 11 SMSR depends mostly on the bias current and the mirror loss margin, Δα: SMSR(dB) ≈ 10 log10 where δ G = (α i + α m (λ0 ) )β spη r ∆α + ∆G δG 12 ∆α + ∆G SMSR( dB) ≈ 10 log10 + 1 δG + 1 1 I / I th − 1 ∆α = αm(λ1) - αm(λ0) is the loss margin, ∆G = Γg(λ 0) - Γg(λ 1) is the modal gain margin and δG is the net modal gain for the main mode, βsp is the spontaneous emission factor and ηr is the radiative recombination efficiency etc. Modal positions and corresponding mirror losses are solved with Transfer Matrix Method Threshold gain condition: Shown for 980 nm grating + alignment with gain spectrum 2 HR AR R1 r (λ , α ) = 1 Phase condition: Both transverse and longitudinal mode discrimination have to be taken into account r(λ,α) has to be real and positive CAS 2010 Tuesday, Oct. 12, O2 higher order transverse modes might have higher coupling coefficients but ... CAS 2010 Tuesday, Oct. 12, O2 Transverse modal selectivity Longitudinal mode selectivity (confinement factor discrimination) (κL) 13 Thickness of the un-etched cladding (nm) ∆α + ∆G SMSR( dB) ≈ 10 log10 + 1 δG ( Γ1 - Γ2 ) / Γ1 500 450 400 350 300 250 200 150 100 50 0.5 1.0 1.5 mirror loss discrimination modal gain discrimination 2 Gm = ∫ ∫ Ψm ( x , y ) ⋅ g m ( x , y ) ∫∫ ( Γ1 - Γ3 ) / Γ1 2.0 2.5 0.5 1.0 1.5 2.0 2 ± ∞ Ψm ( x, y ) dx dy SMSR increases with κL only up to a certain κL value: • small κL-value → not enough modal selectivity • high κL-value → spatial hole burning → multiple longitudinal modes 2.5 Simulated SMSR for a DFB structure with phase-matched facets, R 1 = 95 %, R 2 = 2 %, L = 500 µm and αi = 5, 10, 15 or 20 cm-1 when (a) κ is either 0 or 20 cm-1 and I is varied, and (b) I/Ith = 3 and κ is varied Width of the middle section, W (µm) Thickness of the un-etched cladding (nm) 2 (Γ1-Γ2)*(Γ1-Γ3)/Γ1 500 450 400 350 300 250 200 150 100 50 0.5 1.0 1.5 2.0 0 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00 2.5 14 dxdy => κL≈1-2 gm ( x, y ) act.reg _ under−ridge = g 2 Gm ≈ g ∫ ∫ act reg , under ridge Ψm ( x, y ) ∫∫ Width of the middle section, W (µm) Γ = 1m dx dy = gΓm Ψ 2 ( x, y ) dx dy ±∞ m g Γ − gΓ Γ −Γ m= 1 m 1 gΓ Γ 1 1 κL≈1-2 => 1-2 longitudinal modes within the grating stopband CAS 2010 Tuesday, Oct. 12, O2 CAS 2010 Tuesday, Oct. 12, O2 Coupling coefficient Modal degeneracy (targets & limits) (phase(phase-shift section & mirror reflections) 80 3,205 3,204 70 60 3,203 50 40 3,202 30 20 3,201 10 0 0,0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 Filling factor, γ grating trench width is an essential parameter for achieving high κ 0,9 3,200 1,0 100 Use of end-mirror reflections yellow blue Devices without and with a phase-shift section when end facets have perfect anti-reflection (AR) coatings the SMSR deteriorates rapidly for facet reflectivities of only a few percent due to the phase randomness of the facet reflections K2281LD894 with W=1.5 µm, D=0.5 µm and t=0 nm st 1 transverse mode 90 Coupling coefficient and 80 st nd 1 -order grating, rd 3 -order grating 16 Use of a phase-shift section The product κL=1-2 is important for longitudiunal mode selectivity High κ is critical only when devices have to be short (high modulation bandwidth) When the devices can be long (narrow linewidth) the technological limits matter more Coupling coefficient, κ [1/cm] Coupling coefficient, κ [1/cm] 90 K2281LD894 with W=1.5 µµm, D=0.5 µµm and t=0 nm st 1 transverse mode st nd Coupling coefficient 1 -order grating, 2 -order grating rd 3 -order grating; and Average effective refractive index and Average effective refractive index 15 100 2 -order grating HR 70 Etch aspect ratio > 10 < 10 60 50 3rd-order grating with γ=½ 20 yellow AR / high reflection (HR) –coated device without a phase-shift section when the distance between HR coating and the last grating period is varied => all DFB laser designs have an element of randomness as a result of the uncontrollable phase shift associated with the facet reflections. 40 30 red 10 0 0 50 100 150 200 250 300 Trench width, t [nm] 350 400 450 CAS 2010 Tuesday, Oct. 12, O2 AR CAS 2010 Tuesday, Oct. 12, O2 End-mirror influence Power – linewidth (randomness in laser characteristics) (compromise evaluation) 0.1 0.0 0.0 0.1 0.2 0.3 0.4 0.5 0.0 0.1 0.2 0.3 0.4 0.5 0.0 0.1 0.2 0.3 0.4 0.5 0.5 0.85 0.4 0.95 0.3 1.05 0.2 1.15 0.1 1.25 SMSR (dB) 23.0 25.0 27.0 29.0 31.0 33.0 35.0 37.0 39.0 41.0 (Optical thickness of d1)/λ 1.35 0.0 0.0 0.1 0.2 0.3 0.4 0.5 0.0 0.1 0.2 0.3 0.4 0.5 0.5 Ith (mA) ∆ν (MHz) 0.3 0.2 0.1 0.0 0.0 0.1 0.2 0.3 0.4 0.5 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 SMSR (dB) 0.0 0.1 0.2 0.3 0.4 0.5 improves degrades 21.0 23.3 25.5 27.8 0.0 0.1 0.2 0.3 0.4 0.5 30.0 6.3 improves κκL = 0.0 κκL = 0.5 κκL = 1.0 κκL = 1.5 κκL = 2.0 1 7.1 6.6 9.1 6.9 11.1 0,1 0 13.1 20 15 10 κ, L increase => Δν↓, Pout↓ achieving both Pout>15 mW and Δν < 1.5 MHz while 0.5 < κL < 2 (in order to have high SMSR) possible only in limited κ, L ranges 5 10 20 30 Current (mA) 40 0 50 Calculated spectral linewidth (blue lines) and output power (red lines) variations for 1.2 mm long 894 nm LC-DFB structure with phase-matched R1=0.95 & R2= 0.02. 15.1 17.1 19.1 20.4 8.4 0.0 0.1 0.2 0.3 0.4 0.5 25 P (mW) 8.1 Ith (mA) λ = 894 nm, neff = 3.2, k = 22 cm-1, L = 700 μm, Ibias = 90mA, αeff = 6 & αi = 25 cm-1 R1 = 0.95 & R2 = 0.02 7.8 (Optical thickness of d2 )/λ R 2↑ 18.8 7.5 0.0 0.1 0.2 0.3 0.4 0.5 22 24 26 28 30 32 34 36 38 40 42 44 46 16.5 18 10 14.3 7.2 (Optical thickness of d2)/λ 0.4 12.0 0.0 0.1 0.2 0.3 0.4 0.5 R1 = 0.95 & R2 = 0.275 5.4 6.3 7.2 8.1 9.0 9.9 10.8 11.7 12.6 13.5 14.4 P (mW) 0 3 6 9 12 15 18 21 24 27 30 ⇒ some important parameters are contradictory influenced by mirror phase (e.g. output power, bandwidth & linewidth) 0.0 0.1 0.2 0.3 0.4 0.5 κ ≈ 22 cm-1, L = 550 µm => κL=1.21 simulated Δν≈0.7 MHz, Pout≈20 mW κ ≈ 22 cm-1, L = 1000 µm => κL=2.2 simulated Δν≈0.35 MHz, Pout<10 mW R1 = 0.95 & R2 = 0.95 6 40 -1 30 κ = 10 cm 4 25 -1 κ = 5 cm 3 20 15 2 1 10 κ= -1 20 cm -1 10 cm -1 5 cm 5 0 1000 1500 2000 L ( µm) CAS 2010 Tuesday, Oct. 12, O2 Nanoimprint lithography (NIL) 35 -1 500 degrades 0.5 < κ L < 2.0 κ = 20 cm 5 2500 3000 3500 0 4000 CAS 2010 Tuesday, Oct. 12, O2 NIL stamp structure 19 Contact technique involving pressure (small features => potential for very high pressure) and mechanical deformation => requirements: Power (mW) when I = 50 mA (Optical thickness of d1)/λ (Optical thickness of d2)/λ ∆ν (MHz) P (mW) 6.8 7.3 7.8 8.3 8.8 9.3 9.8 10.3 10.8 11.2 Linewidth (MHz) when I = 50 mA 0.2 Ith (mA) 24.0 26.0 28.0 30.0 32.0 34.0 36.0 38.0 40.0 42.0 44.0 46.0 Increasing kL 0.3 SMSR (dB) 0.97 1.10 1.23 1.36 1.49 1.62 1.75 1.88 2.00 Linewidth (MHz) ∆ν (MHz) 0.4 Power (mW) (Optical thickness of d1)/λ 17 0.5 Stamp should be soft and flexible (to distribute pressure and to accomodate particle impurities and wafer non-flatness) Stamp should have minimal lateral deformation during imprinting (mostly for small features) Applied pressure should be low and uniform, despite varying imprinting area aspect ratio (avoid breaking the wafer and avoid local deformation of small size features – high local pressure) No temperature cycling and stress (=> UV-NIL => stamp should be UV transparent) Resist should have low viscosity – has to flow almost freely (surface tension forces) No bubbles should be present (vacuum process + low viscosity + surface tension) Small residual resist layer after imprint (residual layer has to be removed => smallest feature size is comparable with residual layer thickness) Durability of master template ( contamination) and stamps CAS 2010 Tuesday, Oct. 12, O2 20 Three-layer NIL masks: PDMS (polydimethylsiloxane), which is an UV transparent polymer • • Hard PDMS (h-PDMS) is used for copying the nanopatterns from the master (does not deform –good replication of small features where the local forces are very high - potential high deformation) Soft PDMS (s-PDMS), which is very poor for replicating small, sub 300 nm, patterns, is used as a cushion to ensure stamp softness. Three-layer sequence: • • • • a thick s-PDMS cushion layer (softness, deformation, distribution of pressure), a thin layer of glass (which preserves the stamp flexibility but prevents the lateral stretching) a relatively thin nano-patterned h-PDMS layer. This three-layer soft stamp (B) may be placed on a relatively thick glass plate, generating a hard stamp (A). accommodates particles and non-flat wafer 10 µm CAS 2010 Tuesday, Oct. 12, O2 NIL imprinting Improving the etching profile 21 Previous NIL masks produces thick residual layers (100 nm), which limits grating resolution New mask design with resist escape structures allow a reduction of residual layers below 10 nm below 10 nm features are imprintable 22 lateral leakage control AFM 1µ µm K2281LD894, grating order 1; W = 1.5 µm, t = 0 nm Coupling coefficient of the 1st mode, κ [1/cm] Imprint Coupling c oefficient of the 1st mode, κ [1/cm] Template DeLight New Dilute-Nitride: W = 1.5 µ m D = 2.0 µm t = 0 nm 35 30 25 20 15 10 5 0 0 => etching profile accuracy is essential in achieving high κ 100 200 300 Triangle side length, x [nm] for an ideal etching profile 0,5 1,0 1,5 2,0 2,5 CAS 2010 Tuesday, Oct. 12, O2 23 Narrow trench etching experiments K2281LD894 with W=1.5 µµm, D=0.5 µµm and t=0 nm st 1 transverse mode st nd Coupling coefficient 1 -order grating, 2 -order grating rd and 3 -order grating 90 80 70 Etch aspect ratio 60 50 30 25 20 15 10 5 0 20 10 0 100 150 200 300 400 D=3.0 µm & ideal etching profile rd 3 -order grating with γ=½ 30 50 100 Triangle side length, x [nm] 40 0 24 DeLight New Dilute-Nitride: W = 1.5 µm D = 2.0 µ m t = 0 nm 35 0 > 10 < 10 200 250 300 350 400 Trench width, t [nm] D=0.5 µm & ideal etching profile 450 Coupling coefficient, κ [1/cm] Coupling coefficient, κ [1/cm] Coupling coefficient of the 1st mode, κ [1/cm] 23 100 3,0 Width of the lateral extension, D [µm] The coupling coefficient improvement with extended lateral gratings is derived mainly from improved etching profile and less from the increased lateral extension itself. CAS 2010 Tuesday, Oct. 12, O2 Higher κ by using wider lateral extension of the gratings & narrow trench widths 400 55 50 45 40 35 30 25 20 15 10 5 0 0,0 K2281LD894 with W=1.5 µµm, D=3. 0 µµm and t=0 nm st 1 transverse mode 140 130 120 110 100 90 80 70 60 50 40 30 20 10 0 Coupling coefficient and 0 50 100 => trench widths < 100 nm are desirable irrespective of grating order 150 st 1 -order grating, rd 3 -order grating 200 250 nd 2 -order grating SEM views of gratings with period Λ=180 nm and trench widths of 140 nm (left panel),120 nm (middle panel) and 100 nm (right panel). Gratings were etched to a depth of 1.2, 1.2 and 1.1 µm in GaAs/AlGaAs 300 350 400 450 High aspect-ratio-patterns with BCB planarization: -Better mechanical stability -Less optical losses due scattering -Higher bandwidth Trench width, t [nm] => 100 nm trench widths are achievable CAS 2010 Tuesday, Oct. 12, O2 CAS 2010 Tuesday, Oct. 12, O2 Experimental evaluation of the coupling coefficient Fabricated 894 nm LC-DFB laser characteristics 25 26 -56 8 0,40 2,0 -62 -58 6 Power (mW) Power (dBm) Power (dBm) 7 -60 -60 -62 o 10 C o 20 C o 30 C o 40 C o 50 C o 60 C 0 70 C 0 80 C 5 4 3 2 -64 891 892 893 894 Wavelength (nm) -64 891 895 1 892 3rd-order grating with γ=0.5 893 Wavelength (nm) 894 895 0 0 5 10 Coupling coefficient (1/cm) -54 Power (dBm) -56 -58 -60 -62 891 892 893 Wavelength (nm) 894 Upper error limit Fitted κ Lower error limit 20 25 30 35 40 45 1,0 0,5 0,25 0,20 0,15 0,10 0,05 0,0 50 0,00 cm-1, m = 3, γ = 0.5, κ = 22 L=550 μm, κL=1.21 (simulated / measured) emission at 894.6 nm (tunable ± 2nm) Ith ≈ 10 / 10 mA output power (@ I bias = 90 < 100 mA) > 25 / ≈ 15 mW single mode CW operation with SMSR > 40 / 35 dB spectral linewidth (@ L=1000 µm) < 0.35 / ≈ 1.2 MHz 25 20 15 895 15 0,35 0,30 1,5 Current (mA) 30 Measurement (8.0mA) Fitting by LAPAREX 0,45 2,5 Voltage (V) -58 9 Measurement (7.6mA) Fitting by LAPAREX 7,0 7,5 Current (mA) 8,0 Coupling coefficient extraction by fitting simulated and experimental sub-threshold spectra CAS 2010 Tuesday, Oct. 12, O2 CAS 2010 Tuesday, Oct. 12, O2 Further linewidth reduction Linewidth evaluation @ 894 nm 27 100 MHz AO MOD υ DET υ +100 MHz υ Technion 100 MHz Spectrum analyzer Fiber delay line length L 28 The spectral linewidth, Δν, of a laser due to spontaneous emission noise and carrier noise: spectral linewidth simulation (@ L=1000 µm) < 0.35 MHz DFB Slope (W/A) -54 Measurement (6.9mA) Fitting by LAPAREX Self-homoheterodyne spectrum obtained with a 2 km long fiber (assuming Δν=1 MHz, the coherent length in fiber is ≈ 200 m) Self-homoheterodyne experimental set up ∆ν = ΓR sp' 4πN p 2 ' sp qυ g n (1 + α ) = Γ4Rπηqυ(I −g IV ) (1 + α ) = 4Γπη (α (I − I ) 2 g 2 th act eff g th sp eff i th i th i ( + α m ) 1 + α eff 2 2 ) = 4πηqυ(I n− I g i sp th ) (α i + α m )2 (1 + α eff 2 ) Decreasing αm is more straightforward than decreasing αi and αeff. However, the output power is proportional to αm/(αi +αm). Linewidth and light-current measurements suggest that αeff and αi are higher than evaluated in simulations => significant linewidth reduction can be obtained by the reduction of αi with epilayer and grating design and/or by the reduction of αeff with a phase-shifted DFB design Short-delay self-homodyne experimental setup Short-delay homodyne spectrum fitting (left panel) linewidth evaluation at different bias levels (right panel) for a 1000 µm long LC-DFB laser CAS 2010 Tuesday, Oct. 12, O2 αeff =6 was used in our linewidth simulations CAS 2010 Tuesday, Oct. 12, O2 Acknowledgements 29 Development of Extremely Narrow-Band Semiconductor Distributed Feedback Laser Technology (Narrow DFB) Thank you! CAS 2010 Tuesday, Oct. 12, O2

© Copyright 2020